Dataset Information

Development and validation of a prediction model for active tuberculosis case finding among HIV-negative/unknown populations.

ABSTRACT: A prediction model of prevalent pulmonary tuberculosis (TB) in HIV negative/unknown individuals was developed to assist systematic screening. Data from a large TB screening trial were used. A multivariable logistic regression model was developed in the South African (SA) training dataset, using TB symptoms and risk factors as predictors. The model was converted into a scoring system for risk stratification and was evaluated in separate SA and Zambian validation datasets. The number of TB cases were 355, 176, and 107 in the SA training, SA validation, and Zambian validation datasets respectively. The area under curve (AUC) of the scoring system was 0·68 (95% CI 0·64-0·72) in the SA validation set, compared to prolonged cough (0·58, 95% CI 0·54-0·62) and any TB symptoms (0·6, 95% CI 0·56-0·64). In the Zambian dataset the AUC of the scoring system was 0·66 (95% CI 0·60-0·72). In the cost-effectiveness analysis, the scoring system dominated the conventional strategies. The cost per TB case detected ranged from 429 to 1,848 USD in the SA validation set and from 171 to 10,518 USD in the Zambian dataset. The scoring system may help targeted TB case finding under budget constraints.

SUBMITTER: Shih YJ

PROVIDER: S-EPMC6467872 | biostudies-literature | 2019 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Development and validation of a prediction model for active tuberculosis case finding among HIV-negative/unknown populations.

Shih Yun-Ju YJ Ayles Helen H Lönnroth Knut K Claassens Mareli M Lin Hsien-Ho HH

Scientific reports 20190416 1

A prediction model of prevalent pulmonary tuberculosis (TB) in HIV negative/unknown individuals was developed to assist systematic screening. Data from a large TB screening trial were used. A multivariable logistic regression model was developed in the South African (SA) training dataset, using TB symptoms and risk factors as predictors. The model was converted into a scoring system for risk stratification and was evaluated in separate SA and Zambian validation datasets. The number of TB cases w ...[more]

PMID: 30992463

Similar Datasets

Project description:BackgroundThere is limited evidence on whether active case finding (ACF) among marginalised and vulnerable populations mitigates the financial burden during tuberculosis (TB) diagnosis.ObjectivesTo determine the effect of ACF among marginalised and vulnerable populations on prevalence and inequity of catastrophic costs due to TB diagnosis among TB-affected households when compared with passive case finding (PCF).MethodsIn 18 randomly sampled ACF districts in India, during March 2016 to February 2017, we enrolled all new sputum-smear-positive TB patients detected through ACF and an equal number of randomly selected patients detected through PCF. Direct (medical and non-medical) and indirect costs due to TB diagnosis were collected through patient interviews at their residence. We defined costs due to TB diagnosis as 'catastrophic' if the total costs (direct and indirect) due to TB diagnosis exceeded 20% of annual pre-TB household income. We used concentration curves and indices to assess the extent of inequity.ResultsWhen compared with patients detected through PCF (n = 231), ACF patients (n = 234) incurred lower median total costs (US$ 4.6 and 20.4, p < 0.001). The prevalence of catastrophic costs in ACF and PCF was 10.3 and 11.5% respectively. Adjusted analysis showed that patients detected through ACF had a 32% lower prevalence of catastrophic costs relative to PCF [adjusted prevalence ratio (95% CI): 0.68 (0.69, 0.97)]. The concentration indices (95% CI) for total costs in both ACF [-0.15 (-0.32, 0.11)] and PCF [-0.06 (-0.20, 0.08)] were not significantly different from the line of equality and each other. The concentration indices (95% CI) for catastrophic costs in both ACF [-0.60 (-0.81, -0.39)] and PCF [-0.58 (-0.78, -0.38)] were not significantly different from each other: however, both the curves had a significant distribution among the poorest quintiles.ConclusionACF among marginalised and vulnerable populations reduced total costs and prevalence of catastrophic costs due to TB diagnosis, but could not address inequity.

Project description:Background: Community-based active case finding (ACF) for tuberculosis (TB) implemented among marginalised and vulnerable populations in 285 districts of India resulted in reduction of diagnosis delay and prevalence of catastrophic costs due to TB diagnosis. We were interested to know whether this translated into improved treatment outcomes. Globally, there is limited published literature from marginalised and vulnerable populations on the independent effect of community-based ACF on treatment outcomes when compared to passive case finding (PCF). Objectives: To determine the relative differences in unfavourable treatment outcomes (death, loss-to-follow-up, failure, not evaluated) of ACF and PCF-diagnosed people. Methods: Cohort study involving record reviews and interviews in 18 randomly selected districts. We enrolled all ACF-diagnosed people with new smear-positive pulmonary TB, registered under the national TB programme between March 2016 and February 2017, and an equal number of randomly selected PCF-diagnosed people in the same settings. We used log binomial models to adjust for confounders. Results: Of 572 enrolled, 275 belonged to the ACF and 297 to the PCF group. The proportion of unfavourable outcomes were 10.2% (95% CI: 7.1%, 14.3%) in the ACF and 12.5% (95% CI: 9.2%, 16.7%) in the PCF group (p = 0.468). The association between ACF and unfavourable outcomes remained non-significant after adjusting for confounders available from records [aRR: 0.83 (95% CI: 0.56, 1.21)]. Due to patient non-availability at their residence, interviews were conducted for 465 (81.3%). In the 465 cohort too, there was no association after adjusting for confounders from records and interviews [aRR: 1.05 (95% CI: 0.62, 1.77)]. Conclusion: We did not find significant differences in the treatment outcomes. Due to the wide CIs, studies with larger sample sizes are urgently required. Studies are required to understand how to translate the benefits of ACF to improved treatment outcomes.

Project description:BackgroundCommunity-based active case-finding interventions might identify and treat more people with tuberculosis disease than standard case detection. We aimed to assess whether active case-finding interventions can affect tuberculosis epidemiology in the wider community.MethodsWe did a systematic review by searching PubMed, Embase, Scopus, and Cochrane Library for studies that compared tuberculosis case notification rates, tuberculosis disease prevalence, or tuberculosis infection prevalence or incidence in children, between populations exposed and unexposed to active case-finding interventions. We included studies published in English between Jan 1, 1980, and April 13, 2020. Studies of active case-finding in the general population, in populations perceived to be at high risk for tuberculosis, and in closed settings were included, whereas studies of tuberculosis screening at health-care facilities, among household contacts, or among children only, and studies that screened fewer than 1000 people were excluded. To estimate effectiveness, we extracted or calculated case notification rates, prevalence of tuberculosis disease, and incidence or prevalence of tuberculosis infection in children, and compared ratios of these outcomes between groups that were exposed or not exposed to active case-finding interventions.Results27 883 abstracts were screened and 988 articles underwent full text review. 28 studies contributed data for analysis of tuberculosis case notifications, nine for prevalence of tuberculosis disease, and two for incidence or prevalence of tuberculosis infection in children. In one cluster-randomised trial in South Africa and Zambia, an active case-finding intervention based on community mobilisation and sputum drop-off did not affect tuberculosis prevalence, whereas, in a cluster-randomised trial in Vietnam, an active case-finding intervention based on sputum tuberculosis tests for everyone reduced tuberculosis prevalence in the community. We found inconsistent, low-quality evidence that active case-finding might increase the number of cases of tuberculosis notified in populations with structural risk factors for tuberculosis.InterpretationCommunity-based active case-finding for tuberculosis might be effective in changing tuberculosis epidemiology and thereby improving population health if delivered with high coverage and intensity. If possible, active case-finding projects should incorporate a well designed, robust evaluation to contribute to the evidence base and help elucidate which delivery methods and diagnostic strategies are most effective.FundingWHO Global TB Programme.

Project description:BackgroundInvestigation of household contacts exposed to infectious tuberculosis (TB) is widely recommended by international guidelines to identify secondary cases of TB and limit spread. There is little data to guide the use of contact investigations outside of the household, despite strong evidence that most TB infections occur outside of the home in TB high burden settings. In older adolescents, the majority of infections are estimated to occur in school. Therefore, as part of a project to increase active case finding in Swaziland, we performed school contact investigations following the identification of a student with infectious TB.MethodsThe Butimba Project identified 7 adolescent TB index cases (age 10-20) with microbiologically confirmed disease attending 6 different schools between June 2014 and March 2015. In addition to household contact investigations, Butimba Project staff worked with the Swaziland School Health Programme (SHP) to perform school contact investigations. At 6 school TB screening events, between May and October 2015, selected students underwent voluntary TB screening and those with positive symptom screens provided sputum for TB testing.ResultsAmong 2015 student contacts tested, 177 (9%) screened positive for TB symptoms, 132 (75%) produced a sputum sample, of which zero tested positive for TB. Household contact investigations of the same index cases yielded 40 contacts; 24 (60%) screened positive for symptoms; 19 produced a sputum sample, of which one case was confirmed positive for TB. The odds ratio of developing TB following household vs. school contact exposure was significantly lower (OR 0.0, 95% CI 0.0 to 0.18, P = 0.02) after exposure in school.ConclusionSchool-based contact investigations require further research to establish best practices in TB high burden settings. In this case, a symptom-based screening approach did not identify additional cases of tuberculosis. In comparison, household contact investigations yielded a higher percentage of contacts with positive TB screens and an additional tuberculosis case.

Dataset Information

Development and validation of a prediction model for active tuberculosis case finding among HIV-negative/unknown populations.

Publications

Development and validation of a prediction model for active tuberculosis case finding among HIV-negative/unknown populations.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets