SUMMARY
BACKGROUND:
The yield of TB contact tracing is often limited by challenges in reaching individuals during the screening process. We investigated the times at which index patients and household contacts were typically at home and the potential effects of expanding the timing of home-based contact investigation.
METHODS:
Index patients and household contacts in Kampala, Uganda, were asked about their likely availability at different day/time combinations. We calculated the “participant identification gap” (defined as the proportion of participants who reported being home <50% of the time) during business hours only. We then estimated the incremental reduction in the participant identification gap if hours were expanded to include weekday evenings, Saturdays, and Sundays. Statistical significance was assessed using McNemar’s tests.
RESULTS:
Nearly half of eligible individuals (42% of index patients and 52% of contacts) were not likely to be home during contact investigation conducted only during business hours. Expanding to weekday evenings, Saturdays, and Sundays would reduce this participant identification gap to 15% among index patients and 18% among contacts – while also reducing differences by sex and employment.
CONCLUSIONS:
Expanding hours for conducting contact investigation or other home-based health interventions could substantially reduce the number of individuals missed and address disparities in access to care.
Keywords: epidemiology, contact investigation, prevention/control program, tuberculosis
Of the estimated 10 million new cases of TB annually, nearly 40% are not diagnosed and notified to health authorities.1 It is therefore important to investigate how access to TB diagnostic evaluation might be increased, especially among those at highest risk. Contact investigation, a systematic process of screening and testing for TB among close contacts of index patients,2 is a high-yield approach for identifying individuals with TB in low-income settings.3,4 Initially recommended by the WHO in 2012,5 household contact investigation has been incorporated into national guidelines in many low- and middle-income countries, including Uganda. However, implementation has been inconsistent. A recent meta-analysis found that 3–91% of eligible index cases and 43–100% of eligible contacts completed contact investigation in high-burden settings.6 Identifying barriers to implementation of effective contact investigation can improve both reach and yield.
In Uganda, contact investigation generally relies on the index patient naming individuals with whom they share a household; those individuals are then requested to come to the health center for evaluation.7 Many contacts may not come to the clinic for evaluation for reasons including insufficient time and high travel costs.8 Therefore, home-based contact investigation, in which health workers visit homes of newly diagnosed patients to screen co-habitants for TB and refer individuals to clinics for evaluation and treatment, may increase participation by bringing services to the home.9,10 However, even with home-based screening, one study in Uganda found that only 50% of index cases completed home visits.10 Among household contacts encountered during these visits, only 20% completed recommended evaluation for TB; scheduling and completing the home visit were identified as key barriers.11 Many household members, especially men, are often not home during regular business hours, when contact investigation is commonly conducted.12 Expanding hours for conducting home visits might therefore help to overcome a key barrier to healthcare access and driver of health disparity. We sought to understand whether the timing of home visits might affect the yield of home-based contact investigation for TB.
METHODS
Study overview and enrollment
This is a cross-sectional secondary data analysis nested in a study of TB transmission in a defined community within Kampala, Uganda (STOMP-TB).13,14 We included and interviewed all people diagnosed with TB (i.e., index patients) enrolled in STOMP-TB from February 2019 to August 2021 and household contacts enrolled from February to December 2019. In March 2020, initial COVID-19 restrictions, including the prohibition of non-essential public and private transport, were implemented; in June 2020, these restrictions were relaxed but an evening curfew and limitations on large gatherings continued throughout the study period.
Index TB patient enrollment
Index patients were enrolled in the study from February 2019 to August 2021 through two mechanisms: facility-based diagnosis and community-based case-finding. All consenting adult (≥15 years) residents of the study area who were diagnosed with pulmonary TB at one of the four TB diagnosis and treatment centers located within the study area were enrolled as health facility index patients. Study staff played no role in diagnosing patients at health facilities but recruited eligible participants after completion of their visit. At the main public health facility, we also enrolled adult index patients who resided outside the study area but received TB diagnosis and/or treatment at that facility from January 2020 to February 2021.
We conducted active case-finding campaigns with the goal of collecting and testing sputum from all adult residents of the study area for TB, regardless of symptoms, using Xpert® MTB/RIF Ultra (Cepheid, Sunnyvale, CA, USA) from February to December 2019 and February to August 2021. Study staff conducted sensitization events with local leaders and health workers in each study region. Research staff went door-to-door (7 days a week) throughout the study area asking adult residents for sputum for TB testing. At the same time, 20 scheduled venue-based screening events were held on weekends in public spaces (e.g., churches and markets). Screening activities primarily took place between the hours of 8 am and 5 pm. All screening participants with positive (including trace) Xpert results were recruited as community-based index patients and referred to their local health facility for treatment.
Contact enrollment
From February to December 2019, all enrolled index patients who resided in the study area were asked to name their household contacts of any age (those who routinely spend at least one night a week in the household) for TB evaluation. Study staff then made up to three attempts to screen and enroll the named contacts, including scheduling times in advance, offering screening on weekends, and offering screening at location(s) convenient to the participant (work, home, other venues, etc.). Consenting contacts were screened for TB symptoms, asked to provide sputum for Ultra, and offered the tuberculin skin test (TST); in the event that a contact was diagnosed with TB, they were then enrolled as a community-based index patient. Participants with a positive TST were referred to local facilities for follow-up; preventive therapy was offered to children <5 years old and participants with HIV per the Uganda National TB Guidelines.15
Analysis
The primary outcome of interest was self-reported likelihood of being at home at a given time for home-based contact investigation. Index TB patients and household contacts were asked how often they were typically at home (for example, in each week during the last month) at three specific time points (9 am, 3 pm, and 7 pm) on weekdays, Saturdays, and Sundays. For each of these times, participants were asked to estimate the proportion of the time they were likely to be home: rarely (<20% of the time), sometimes (20–49%), often (50–80%), or always (>80%) (Supplemental Table S1). For this analysis, we defined those who were at home ≥50% of the time as “likely to be home” or “typically available.” In a sensitivity analysis, we compared this definition to the weighted mean proportion of index patients and contacts who were likely to be home at each day/time. To calculate the weighted mean, we multiplied the midpoint of the range (i.e., assuming that 10% of individuals who reported being home 0–20% of the time would be home) by the number of individuals who reported that answer and summed this across the four different groups (sometimes, rarely, often, always).
We first considered a scenario in which contact investigation was conducted during routine business hours only, assuming that this approach would reach participants self-reporting as typically being home on weekdays at 9 am or 3 pm. We calculated the “participant identification gap” as the proportion of participants who were unlikely to be home at both 9 am and 3 pm under this scenario. We then quantified the reduction in the identification gap (both overall and stratified by sociodemographic characteristics) that could be achieved by expanding the times of conducting contact investigation. In doing so, we assumed patients or contacts who self-reported being typically at home during at least one proposed time would be reached. For example, any patients or contacts who reported typically being home on weekdays at 9 am and/or 3 pm would be considered reached by the “business hours only” contact investigation; an individual who reported typically not being home at both these times would be considered missed – but if that person reported being typically at home on a Saturdays at 9 am, they would be considered reached under expanded Saturday hours. We compared the participant identification gap under the “business hours only” scenario to the most inclusive scenario (business hours, weekday evenings, Saturdays, and Sundays) using McNemar’s tests.
Ethical considerations
The study was approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board, Baltimore, MD, USA (IRB Number 11353) and the Higher Degrees, Research and Ethics Committee of the Makerere University School of Public Health, Kampala, Uganda (Study Protocol Number 544). All index patients and contacts provided written informed consent (or written assent and parental consent for those <18 years old) for study activities.
RESULTS
Study population
We surveyed 598 participants (Table 1), including 385 index patients (218 [57%] from health facilities and 167 [43%] from community-wide screening) and 213 household contacts (of 559 contacts named by index patients). Two of the 167 community index patients were identified through contact investigation (Supplementary Figure S1). Only eight index patients were eligible for the study but did not enroll. There was no significant difference in age or sex between the contacts who were enrolled and those who could not be surveyed. Compared to index patients, contacts were younger (median age: 17 years, interquartile range [IQR] 7–29 vs. 31 years, IQR 25–40) and less likely to be HIV-positive (7% vs. 27% for index patients); other characteristics were similar (Table 1). The majority of both index patients (62%) and contacts (55%) were male; among those ≥15 years old, lack of formal education (33% of index patients; 34% of adult contacts) and unemployment (25% of index patients, 29% of adult contacts) were common. Among those employed, only 18% of index patients and 29% of contacts reported salaried work, while the remainder were self-employed or worked piece jobs.
Table.
Characteristics of the study population
| Patients with TB n (%) | Household contacts n (%) | Total n (%) | |
|---|---|---|---|
| Total, n | 385* | 213 | |
| Age, years | |||
| 0-14 | 0 | 111 (52) | 111 (19) |
| 15-24 | 89 (23) | 43 (20) | 132 (22) |
| 25-34 | 141 (37) | 24 (11) | 165 (28) |
| 35-44 | 99 (26) | 19 (9) | 118 (20) |
| ≥45 | 56 (15) | 16 (8) | 72 (12) |
| Male sex | 238 (62) | 114 (54) | 352 (59) |
| Highest level of education attained† | |||
| No formal education | 128 (33) | 39 (38) | 167 (34) |
| Primary education | 120 (31) | 33 (32) | 153 (31) |
| Secondary education | 102 (27) | 28 (28) | 130 (27) |
| More than secondary education | 35 (9) | 2 (2) | 37 (8) |
| Unemployed† | 97 (25) | 33 (32) | 130 (22) |
| Parish of residence‡ | |||
| Kisugu | 96 (25) | 84 (39) | 180 (30) |
| Wabigalo | 79 (21) | 61 (29) | 140 (23) |
| Bukasa | 92 (24) | 66 (31) | 158 (26) |
| Other | 118 (31) | 2 (1) | 120 (20) |
| Alcohol use† | |||
| Does not drink alcohol | 183 (48) | 53 (52) | 236 (48) |
| Drinks monthly but not weekly | 119 (31) | 21 (21) | 140 (29) |
| Drinks weekly or more | 83 (22) | 28 (27) | 111 (23) |
| Currently smokes tobacco† | 71 (18) | 10 (10) | 81 (17) |
| HIV-positive | 105 (27) | 13 (7) | 118 (21) |
| Reported any medical condition§ | 17 (6) | 17 (9) | 34 (7) |
Three facility-based index patients were not asked the interview questions about times available at home and were excluded from this analysis.
Contacts aged <15 years were not asked questions pertaining to education, employment, alcohol use, or tobacco (the denominator for these questions is 102 contacts age ≥15 years).
The study area included Kisugu, Wabigalo, and a portion of Bukasa parish; index patients residing outside the study area were ineligible for contact investigation.
Medical conditions included obstructive lung disease, diabetes, high blood pressure, pregnancy, ulcers, back pain, mental problem, throat problem.
Self-reported availability time
At any given time, the proportion of individuals likely to be home (and potentially available for home-based contact investigation) ranged from 47% to 70% for index patients and 42% to 71% for contacts (Figure 1); 85% of index patients and 83% of contacts reported typically being home during at least one time during the week (Supplementary Figure S2). The proportion of index patients and contacts typically at home at each time was similar whether it was estimated as the proportion reporting being at home ≥50% of the time or as a weighted mean. (Supplementary Table S2). The proportion of participants typically at home was similar for index patients enrolled in health facilities vs. the community (Supplementary Table S3). More specific patterns of times at home reported by participants are provided in Supplementary Figure S2.
Figure 1.
Self-reported times at home of the study population. The bars show the percentage of index patients with TB (n = 385, in grey) and household contacts (n = 213, in blue) who reported typically being home (defined as home ≥50% of the time) at each of nine day/time combinations, as shown on the x-axis.
For an intervention conducted during business hours only, 58% of index patients and 48% of contacts were likely to be home. Attempting contact investigation during weekday evenings in addition to business hours increased the proportion likely to be found at home to 75% of index patients and 72% of contacts, and therefore reduced the participant identification gap (i.e., proportion of patients who would be typically not be at home during hours of investigation) from 42% to 25% among index patients and from 52% to 29% among contacts (Figure 2). Further expansion to include weekends could reduce this gap further, to 15% among patients and 18% among contacts. Compared to the baseline scenario of business hours only, each expansion of intervention hours incrementally to include evenings, Saturdays, and Sundays significantly reduced the contact identification gap (P < 0.05 for all comparisons). Expanding hours for conducting contact investigation had a disproportionate effect on closing the participant identification gap among male index patients (56% not likely to be home during business hours vs. 20% with the most inclusive hours), employed individuals with TB (53% vs. 17%), and male contacts (61% vs. 18%) (Figure 3).
Figure 2.
Incremental decrease in TB participant identification gap with expansion of intervention hours. The lines show the participant identification gap – defined as the percentage of individuals who would be missed (i.e., reported being unlikely to be home) by home-based contact investigation – if those activities were restricted to the times described on the x-axis. The initial scenario is contact investigation during business hours (weekdays at 9 am and 3 pm). We then added weekday evenings (7 pm) to the initial business hours-only scenario. The third scenario further adds contact investigation conducted on Saturdays (morning, afternoon, and evening) to the previous scenario of business hours and weekday evenings. The final scenario adds Sundays (morning, afternoon, and evening), thus depicting an intervention where contact investigation would be conducted morning, afternoon, and evening 7 days a week. Each additional time of contact investigation incrementally reduces the corresponding gap in detection. Vertical bars indicate 95% binomial confidence intervals.
Figure 3.
Incremental reduction in TB participant identification gap according to sociodemographic characteristics. The lines show the proportion of individuals who would be missed (i.e., reported being unlikely to be home) by household-based contact investigations conducted at certain times, stratified by sex, age, and employment status. The initial scenario is contact investigation during business hours (weekdays at 9 am and 3 pm). We then added weekday evenings (7 pm) to the initial business hours-only scenario. The third scenario further adds contact investigation conducted on Saturdays (morning, afternoon, and evening) to the previous scenario of business hours and weekday evenings. The final scenario adds Sundays (morning, afternoon, and evening), thus depicting an intervention where contact investigation would be conducted morning, afternoon, and evening seven days a week. Adding weekday evenings, Saturdays, and Sundays would incrementally reduce the corresponding gap in detection.
DISCUSSION
Contact investigation can be a promising approach to expand access to TB diagnostic services and identify individuals with undiagnosed TB.6 However, the effectiveness of contact investigation is often limited by the inability to reach index patients and their contacts for further evaluation.8,11,16 Our analysis of the timing of contact investigation in Kampala, Uganda, suggests that expanding the times during which home-based contact investigation is offered can cut the proportion of individuals not likely to be available (“participant identification gap”) by half or more, with disproportionate impact seen among harder-to-reach sociodemographic groups.
Based on self-reported availability, nearly half of TB patients and contacts would be missed if home-based contact investigations were only conducted during business hours. Men were more likely to be missed during business hours than women. Men are at higher risk for TB,17 and less likely to be connected to the healthcare system than women.18 Increasing the accessibility of TB services for men is therefore a high priority, and our results suggest that expanding hours to include evenings, Saturdays, and Sundays – or scheduling appointments in advance – could improve the proportion of men reached by home-based screening.
Employed individuals were also more likely to be missed during business hours than unemployed individuals. A previous study in Uganda found that contacts with below-median household income or with lower household educational status are less likely to complete TB evaluation,19 suggesting that different groups are likely to experience barriers at different stages of the contact investigation cascade. While unemployment and underemployment are often associated with increased risk of TB,20 employed individuals often face challenges accessing care due to inflexible work schedules and weak employment rights.21 Offering home-based screening during weekends, Saturdays, and Sundays may therefore also help increase access to care for employed individuals.
While our analysis focused on TB contact investigation, our findings also apply to other home-based health interventions. Bringing additional health services to the household represents an opportunity to connect the community with healthcare systems and can also leverage the social influence of the household to increase participation.22 For example, home-based contact investigation for TB could also be utilized as an opportunity to offer HIV testing,23 which has been successfully integrated with screening for non-communicable diseases.24 Furthermore, common schedules of community members could be used to guide implementation of home-based screening and may improve participation in door-to-door TB testing.13 Therefore, an increased understanding when certain sociodemographic populations are likely to be at home is important in planning any home-based health intervention.
Our study had important limitations. First, we relied on self-reported data, summarizing responses as likely to be home if participants reported being home ≥50% of the time. Our sensitivity analysis shows that this definition is robust, but these results may oversimplify actual availability; for example, individuals who are not home most days might still be able to make themselves available on a specified day if an appointment is arranged. Second, it was conducted in an urban setting in Uganda and may not reflect the availability patterns of rural populations. Our findings may nevertheless be generalized to other urban high TB burden settings in sub-Saharan Africa with similar demographics and healthcare systems. Further research collecting data on when individuals are typically at home in other settings or directly testing the impact of offering contact intervention at different times could help confirm our results. Third, by including only those individuals who were available to participate in the study, we may have introduced selection bias; in particular, this may have caused us to underestimate the extent to which likelihood of being home varies at different times, if those with more restrictive schedules were less likely to participate. However, our study team worked continuously in the community, and therefore may have been able to reach more participants than a typical programmatic contact investigation. Fourth, we measured likelihood of being home by self-report and did not confirm responses through direct observation at multiple days and times. Additionally, other factors, such as weather, were not captured in our results, and may affect actual availability. Finally, we focused only on timing and did not estimate costs, human resource limitations, or other patient-level barriers to expanding hours for home-based health services.
In conclusion, this analysis represents one of the first quantitative studies to analyze the factors influencing when individuals are likely to be home and potentially able to participate in home-based contact investigation in urban sub-Saharan Africa. We found that offering contact investigation at additional times could reduce the contact identification gap by more than half, with disproportionate effects among harder-to-reach populations. These findings could be useful for local TB control programs in developing more accessible and equitable approaches to home-based contact investigation and home-based health interventions more broadly.
Supplementary Material
Acknowledgements
The authors thank the STOMP-TB field team for their efforts in case finding and data collection. We also thank the staff and patients at Kisugu Health Center (Kisugu), Alive Medical Services (Kampala), International Hospital Kampala (Kampala), and Meeting Point Clinic (Kampala, Uganda) for their participation, as well as the community members in our study area for their participation and support. This work was made possible through grants provided by the National Institutes of Health, Bethesda, MD, USA (R01HL138728 to DWD and F32HL158019 to KOR).
Footnotes
Conflicts of interest
None declared
References
- 1.World Health Organization. Global tuberculosis report, 2021. Geneva, Switzerland: WHO, 2021. [Google Scholar]
- 2.Fair E, et al. Tuberculosis contact investigation in low- and middle-income countries: standardized definitions and indicators. Int J Tuberc Lung Dis 2015;19(3):269–272. [DOI] [PubMed] [Google Scholar]
- 3.Velleca M, et al. The yield of tuberculosis contact investigation in low- and middle-income settings: a systematic review and meta-analysis. BMC Infect Dis 2021;21(1):1011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Fox GJ, et al. Contact investigation for tuberculosis: a systematic review and meta-analysis. Eur Respir J 2013;41(1):140–156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.World Health Organization. Recommendations for investigating contacts of persons with infectious tuberculosis in low- and middle-income countries. Geneva, Switzerland: WHO, 2012. [PubMed] [Google Scholar]
- 6.Blok L, et al. Comparative meta-analysis of tuberculosis contact investigation interventions in eleven high burden countries. PLoS One 2015;10(3):e0119822. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Davis JL, et al. Home-based tuberculosis contact investigation in Uganda: a household randomised trial. ERJ Open Res 2019;5(3):00112–2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ayakaka I, et al. Identifying barriers to and facilitators of tuberculosis contact investigation in Kampala, Uganda: a behavioral approach. Implemen Sci 2017;12(1):33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kigozi NG, Heunis JC, Engelbrecht MC. Yield of systematic household contact investigation for tuberculosis in a high-burden metropolitan district of South Africa. BMC Public Health 2019;19(1):867. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Armstrong-Hough M, et al. “Something so hard”: a mixed-methods study of home sputum collection for tuberculosis contact investigation in Uganda. Int J Tuberc Lung Dis 2018;22(10):1152–1159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Armstrong-Hough M, et al. Drop-out from the tuberculosis contact investigation cascade in a routine public health setting in urban Uganda: a prospective, multi-center study. PLoS One 2017;12(11):e0187145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dinkelman T, Ngai LR. Time use and gender in africa in times of structural transformation. J Econ Persp 2022;36(1):57–80. [Google Scholar]
- 13.Kendall EA, et al. The spectrum of tuberculosis disease in an urban ugandan community and its health facilities. Clin Infect Dis. 2021. Jun 15;72(12):e1035–e1043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Robsky KO, et al. Spatial distribution of people diagnosed with tuberculosis through routine and active case finding: a community-based study in Kampala, Uganda. Infect Dis Pov 2020;9(1):73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Republic of Uganda Ministry of Health. Manual of the national tuberculosis and leprosy program. 3rded. Kampala, Uganda: MoH, 2017. [Google Scholar]
- 16.Fox GJ, et al. Barriers to adherence with tuberculosis contact investigation in six provinces of Vietnam: a nested case–control study. BMC Infect Dis 2015;15(1):103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Horton KC, et al. Sex differences in tuberculosis burden and notifications in low- and middle-income countries: a systematic review and meta-analysis. PLoS Med 2016;13(9):e1002119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Miller C, et al. ‘It makes the patient’s spirit weaker’: tuberculosis stigma and gender interaction in Dar es Salaam, Tanzania. Int J Tuberc Lung Dis 2017;21(11):S42–S48. [DOI] [PubMed] [Google Scholar]
- 19.Shelby T, et al. Social determinants of tuberculosis evaluation among household contacts: a secondary analysis. Public Health Action 2018;8(3):118–123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Odone A, et al. Association between socioeconomic position and tuberculosis in a large population-based study in rural Malawi. PLoS One 2013;8(10):e77740. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Phiri MM, et al. Improving pathways to care through interventions cocreated with communities: a qualitative investigation of men’s barriers to tuberculosis care-seeking in an informal settlement in Blantyre, Malawi. BMJ Open 2021;11(6):e044944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Armstrong-Hough M, et al. Brief Report: “Give me some time”: facilitators of and barriers to uptake of home-based HIV testing during household contact investigation for tuberculosis in Kampala, Uganda. J Acquir Immune Defic Syndr 2018;77(4):400–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ochom E, et al. Integrating home HIV counselling and testing into household TB contact investigation: a mixed-methods study. Public Health Action 2018;8(2):72–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chamie G, et al. Leveraging rapid community-based HIV testing campaigns for non-communicable diseases in rural Uganda. PLoS One 2012;7(8):e43400. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.