Abstract
SETTING:
Five urban TB diagnostic centres in Lusaka, Zambia.
OBJECTIVE:
To determine the proportion of bacteriologically confirmed pre-treatment lost to follow-up (LTFU) patients with TB found at the study settings in 2020.
DESIGN:
This was a retrospective cohort study in which the TB laboratory and treatment registers at the study sites were cross-matched.
RESULTS:
A total of 1,085 bacteriologically confirmed patients with TB were found in the laboratory TB registers at the study settings. Of these, 809 (74.6%) were males, whereas 8 (0.7%) were children, 1,005 (92.6%) were diagnosed using Xpert, 78 (7.2%) by microscopy. A total of 91 (8.4%, 95% CI 6.8–10.2) were determined to be pre-treatment LTFU. Those who had very low (14.0%, 95% CI 8.5–21.2) and low (11.8%, 95% CI 8.4–16.0) results on Xpert were respectively 3.1 (95% CI 1.6–6.0) and 2.6 (95% CI 1.4–4.8) times more likely to become pre-treatment LTFU than those who had medium (4.5%, 95% CI 2.5–7.4) results. The proportions of pre-treatment LTFU varied among the study sites from 0.7% to 16.1%.
CONCLUSION:
Health facilities should strive to account for every patient with TB, with particular attention to those who are bacteriologically confirmed.
Keywords: bacteriologically confirmed, treatment, cross-match
Abstract
CADRE :
Cinq centres urbains de diagnostic de la TB à Lusaka, en Zambie.
OBJECTIF :
Déterminer la proportion de patients atteints de TB perdus de vue (LTFU) avant le traitement et confirmés bactériologiquement dans les centres d’étude en 2020.
MÉTHODE :
Il s’agit d’une étude de cohorte rétrospective dans laquelle les registres de laboratoire et de traitement de la TB des sites étudiés ont été appariés.
RÉSULTATS :
Au total, 1 085 patients atteints de TB confirmée par une analyse bactériologique ont été trouvés dans les registres de laboratoire de la TB des sites de l’étude. Parmi eux, 809 (74,6%) étaient des hommes, tandis que 8 (0,7%) étaient des enfants. 1 005 (92,6%) ont été diagnostiqués à l’aide d’Xpert, 78 (7,2%) par microscopie. Au total, 91 (8,4% ; IC 95% 6,8–10,2) ont été déterminés comme étant des cas de LTFU avant traitement. Ceux qui avaient obtenu des résultats très faibles (14,0% ; IC 95% 8,5–21,2) et faibles (11,8% ; IC 95% 8,4–16,0) au test Xpert étaient respectivement 3,1 (IC 95% 1,6–6,0) et 2,6 (IC 95% 1,4–4,8) fois plus susceptibles de devenir des cas de LTFU avant traitement que ceux qui avaient obtenu des résultats moyens (4,5% ; IC 95% 2,5–7,4). Les proportions de LTFU avant traitement variaient entre les sites de l’étude de 0,7% à 16,1%.
CONCLUSION :
Les établissements de santé devraient s’efforcer de comptabiliser tous les patients atteints de TB, en accordant une attention particulière à ceux dont la présence est confirmée par la bactériologie.
With a TB incidence of 321 per 100,000 population and HIV prevalence of 11.1% in 2020, Zambia is one of the 30 countries with the highest per capita burden of TB and HIV.1 The notification rate and absolute number of TB cases have been declining over the last decade.2 However, according to WHO estimates, 19,000 people who developed active TB remained undiagnosed.1
Undiagnosed and untreated TB cases are key factors contributing to the continued global TB epidemic, perpetuating TB transmission and increased risk for adverse outcomes due to delayed diagnosis or initiation of treatment.3–5 Finding the missing TB cases is thus a global priority for TB control.6 A systematic review revealed a high (18%) frequency of pre-treatment loss to follow-up in sub-Saharan Africa.7 The review also concluded there was a paucity of evidence on the magnitude of pre-treatment loss to follow-up and recommended urgent evaluation of this problem.
In Zambia, the magnitude of pre-treatment lost to follow-up (LTFU) patients with TB is largely unknown: to date, only one study has been conducted in Chongwe, a rural district. The study reported a pre-treatment loss to follow-up rate of up to 33% among bacteriologically confirmed patients with TB in 2017.8
The aim of the present study was to determine the proportion of pre-treatment LTFU patients with TB overall and by month, facility, sex and age group and mode of diagnosis at five urban TB diagnostic centres in Lusaka, Zambia.
STUDY POPULATION, DESIGN AND METHODOLOGY
Role of laboratories, and diagnostic and treatment centres
The role of TB diagnostic centres in relation to TB control is to provide TB diagnostic services and these sites also serve as TB treatment centres. Laboratories at diagnostic centres receive TB specimens such as sputum and urine from various departments within the health facilities. They also receive specimens from nearest TB treatment centres which do not have the capacity to diagnose TB. At the time of specimen submission, patients’ residential addresses and phone numbers are noted and they are told to return within a week to check for their results.
Laboratory procedure for TB case-finding in Zambia
In Zambia, the Ministry of Health recommends the use of GeneXpert (Cepheid, Sunnyvale, CA, USA) machines for TB diagnosis, while microscopy is used for monitoring treatment for patients who are already on treatment.2 However, in particular situations, such as non-availability or non-functionality of Xpert machines, microscopy may be used for diagnostic purposes. Xpert test results are usually processed in less than 2 h, while those for microscopy take a little longer.
Procedures for the commencement of TB treatment once laboratory results are available
When results are ready at the laboratory, they are dispatched to the chest clinics and referring treatment centres. For those found with TB, chest clinic nurses can either contact them by phone or send TB treatment supporters to follow them up in the community in order to start treatment. Some patients may come on their own to check on the results.
Study sites, populations
The study was conducted at five urban TB diagnostic centres in Lusaka, Zambia. The sites comprised one General Hospital (Chipata General Hospital, population: 494,670), one Zonal Health Centre (Chelstone Zonal Health Centre, population: 136,056) and three urban clinics (Kalingalinga Clinic, population: 103,721; Mtendere Clinic, population: 88,922; Ng’ombe Clinic, population: 75,332). All the five sites had GeneXpert machines and microscopes.
Definitions, study populations and study period
A pre-treatment LTFU TB patient was defined as a bacteriologically confirmed TB patient who did not start TB treatment within 1 month of TB diagnosis. The study population consisted of bacteriologically confirmed patients with TB who were registered at the five health facilities for the period January–December 2020. The study was conducted from December 2021 to June 2022.
Study design and data collecting procedure
This is a retrospective cohort study involving the review of TB registers without direct interaction with TB patients. First, the researchers extracted data from laboratory TB registers at the five health facilities on name, address, sex, age, mode of diagnosis (microscopy, Xpert or lipoarabinomannan [LAM]), name of facility referring sputum samples, date of the laboratory examination, laboratory serial number, specimen examination results, result grades and date treatment was commenced for those who started treatment among bacteriologically confirmed patients with TB. Data were recorded on data collection sheets, and later transferred and stored in an MS Excel file (MicroSoft, Redmond, WA, USA) as a line list.
Data comparison with TB registers
Next, the patient details collected from laboratory registers were cross-matched with those in the TB treatment registers at the chest clinics within the health facilities. This activity was extended to all surrounding clinics which referred sputum samples to diagnostic centres. Patients with matching details in the laboratory and treatment registers and confirmed date of commencement of treatment in the treatment register were believed to have commenced TB treatment. For every matched patient detail, the date of commencement of treatment was copied onto the line list and the name crossed out.
Inclusion criteria
Patients whose names figured in the laboratory TB registers but not in TB treatment registers at chest clinics were considered to be pre-treatment LTFU. Those whose names were found in TB treatment registers but without proof of having started TB treatment (absence of commencement or drug refill dates) were also regarded as pre-treatment LTFU.
Exclusion criteria
Patients who were already on treatment but submitting sputum for monitoring purposes, as well as those patients whose reasons for sputum examination was not indicated in the laboratory register were excluded from the study. Patients with drug-resistant TB were also excluded from the study due to logistical reasons. Patients who died shortly after submitting sputum samples and those who refused to start TB treatment were also excluded.
Analytical method
The proportion of the pre-treatment LTFU patients was calculated as the number of the pre-treatment LTFU patients in the line list divided by the total number of eligible bacteriologically confirmed patients with TB in the study period. The proportion of pre-treatment LTFU patients were compared by month, facility, sex and age group, mode of diagnosis (Xpert or microscopy) and sputum grade.
Statistical method
Statistical analysis was conducted using R software version x64 4.0.2 (R Project for Statistical Computing, Vienna, Austria). Fisher’s exact tests were employed for comparisons of proportions. P < 0.05 was considered statistically significant.
Ethics
This study was conducted as part of the review and monitoring activities of routine TB case finding in relation to TB control. The investigators sought research permission from the Zambia National Health Research Authority, Lusaka, Zambia (Ref No: NHRA 000009/25/11/2021) and ethical clearance from the Institutional Review Board of Research Ethics and Science (ERES) Converge, Lusaka, Zambia. (Ref. No. 2021-Sept-011). Confidentiality was observed and maintained at all stages of the study, and only investigators had access to the collected data.
RESULTS
A total of 1,117 bacteriologically confirmed patients with TB were registered at the five laboratories during the study period. Of these, 21 (1.9%) were diagnosed with drug-resistant TB and excluded from the study due to logistical reasons. A further 10 (0.9%) were excluded because they died shortly after submitting samples for examination and one (0.1%) was excluded for refusing to start TB treatment. The total number of eligible patients was 1,085, 113 of whom were from treatment sites (referred samples from treatment sites). Of the 1,085 eligible patients, 1,005 (92.6%) were diagnosed using Xpert, 78 (7.2%) by microscopy and 2 (0.2%) using rapid urine LAM. Age ranged from 3 to 95 years, 45 (4.1%) had no age indicated in the registers; 809 (74.6%) were males, while 8 (0.7%) were aged <15 years.
A total of 91 (8.4%, 95% confidence interval [CI] 6.8–10.2) were not linked to treatment; hence, they were declared pre-treatment LTFU patients. The Table shows the number and proportion of confirmed patients with TB, those who started treatment and pre-treatment LTFU patients by facility. There were significant differences in pre-treatment loss to follow-up among the facilities. For example, if presumptive patients with TB were found to be bacteriologically positive at Kalingalinga and Chipata had respectively 23.4 (95% CI 3.2–169.5) and 18.2 (95% CI 2.5–130.6) times higher risk of pre-treatment loss to follow-up than Ng’ombe. The number of LTFU patients from referred samples (9/113, 8.0%; 95% CI 3.7–14.6) and those providing samples from diagnostic centres (82/972, 8.4%; 95% CI 6.8–10.4) were not statistically significant (P = 1.0).
TABLE.
Number and proportion of bacteriologically confirmed patients with TB who started treatment and those who did not start treatment (pre-treatment LTFU) by facility, Lusaka, Zambia, 2020
| Facility name | Bacteriologically confirmed patients with TB | Started on treatment n (%) (95% CI) | Pre-treatment LTFU patients n (%) (95% CI) |
|---|---|---|---|
| Chipata General Hospital | 359 (100.0) | 314 (87.5) (83.6–90.7) | 45 (12.5) (9.4–16.3) |
| Chelstone Zonal Clinic | 263 (100.0) | 255 (97.0) (94.1–98.7) | 8 (3.0) (1.3–5.9) |
| Mtendere Clinic | 126 (100.0) | 120 (95.2) (90.0–98.2) | 6 (4.8) (1.8–10.8) |
| Ng’ombe Clinic | 145 (100.0) | 144 (99.3) (96.2–100.0) | 1 (0.7) (0.0–3.8) |
| Kalingalinga Clinic | 192 (100.0) | 161 (83.9) (77.9–88.8) | 31 (16.1) (11.2–22.1) |
| Total, n (%) | 1,085 (100.0) | 994 (91.6) (89.8–93.2) | 91 (8.4) (6.8–10.2) |
LTFU = lost to follow-up; CI = confidence interval.
Figure 1 shows the numbers and the proportions of pre-treatment LTFU patients by sex and age group. Overall, more males (n = 67, 73.6%; 95% CI 63.3–82.3) were pre-treatment LTFU with a peak in the 35–44 years age category. Proportions in the various sex and the age groups were not statistically significantly different (P = 0.16). Figure 2 shows the proportions of the pre-treatment LTFU patients by month. The proportions of pre-treatment LTFU patients among those who were diagnosed using microscopy (7/78, 9.0%; 95% CI 3.7–17.6) and those diagnosed using Xpert (84/1,005, 8.4%; 95% CI 6.7–10.2) were not statistically significant (P = 1.0).
FIGURE 1.
A) Proportions and B) numbers of PT-LTFU patients by sex and age group, Lusaka, Zambia, 2020. Bars indicate 95% confidence intervals. PT-LTFU = pre-treatment lost to follow-up.
FIGURE 2.
Proportion of PT-LTFU patients by month, Lusaka, Zambia, 2020. Bars indicate 95% confidence intervals. PT-LTFU = pre-treatment lost to follow-up.
Figure 3 shows the proportion of pre-treatment LTFU patients by sputum grade. There was no statistical difference by grade among those who were diagnosed using microscopy; however, there were statistically significant differences by grade among those who were diagnosed using Xpert. Those who had very low (14.0%, 95% CI 8.5–21.2) and low (11.8%, 95% CI 8.4–16.0) results on Xpert were respectively 3.1 (95% CI 1.6–6.0) and 2.6 (95% CI 1.4–4.8) times more likely to become pre-treatment LTFU than those who had medium (4.5%, 95% CI 2.5–7.4) results (P = 0.001 and P= 0.001, respectively).
FIGURE 3.
Proportion of PT-LTFU patients by grade of sputum test results, Lusaka, Zambia, 2020. *P < 0.05. †Actual number of Mycobacterium tuberculosis seen per 100 fields under microscope. PT-LTFU = pre-treatment lost to follow-up; AFB = acid-fast bacilli.
DISCUSSION
We conducted a study on pre-treatment loss to follow-up at five TB diagnostic centres in Lusaka, Zambia, from January to December 2020, and found that 8.4% of the eligible patients were lost before commencing treatment; this was almost four times smaller than that found in rural areas of Zambia.8 However, there were significant differences in the proportion of pre-treatment LTFU patients among the facilities studied. There were also significant differences by Xpert sputum grade. Among those who were diagnosed using Xpert, those who had lower positives in their sputum were 2.6–3.1 times more likely to be lost before commencing treatment than those who had higher positives. There was no difference in the proportion of the pre-treatment LTFU patients between those who came from catchment areas of diagnostic centres and those who were referred by treatment centres.
Why we got these findings
There are several reasons why bacteriologically confirmed patients are lost before they start TB treatment, including the long turnaround time (TAT) for sputum results. For example, a study conducted at two of the diagnostic sites—Chelstone Zonal Clinic and Chipata General Hospital in Lusaka—showed that TAT for sputum results involving treatment centres was generally long, ranging from a few days to as long as 3 weeks.9 The prolonged TAT was due to sample overload compared to the limited number of Xpert machines, delay in examining samples and dispatching the results by laboratory personnel and delay in collecting results from the laboratory.9 The consequence of long TATs is that some patients are lost as they get tired of following up their results. Health facilities should therefore work towards reducing TAT.
Another factor that could have caused the pre-treatment loss to follow-up was probably the breakdown in communication between the patients and healthcare workers. One study confirmed that the lack of proper communication between patient and healthcare providers with regard to next steps leads to loss of patients before initiating treatment.10 As explained by health workers at the study sites, breakdown in communication is sometimes due to incorrect phone numbers, physical addresses and change of residence without informing health personnel, which made it difficult for such patients to be followed up in the community.
Some patients may be lost because they do not want to return to health facilities for various personal and healthcare-related reasons. As well-informed patients are more likely to make good choices, patient education, including proper counselling at first contact, may help in addressing this issue.
Kalingalinga and Chipata recorded higher rates of pre-treatment loss to follow-up. This could be as a result of a weak linkage between the laboratories and the chest clinics, long TAT, shortcomings in the collection of patient contact details and inadequate patient education. To improve the situation, there should be a strong linkage between the laboratories and chest clinics, prompt processing of specimens, adequate patient counselling and accurate collection of patient contact details.
Patients with TB who had lower positives in their sputum were more likely to be lost before commencing treatment probably because their symptoms and illnesses were not as severe as those who had higher positives and they may not have expected to be TB-positive.11
Comparisons with the past studies
Our study findings are consistent with the findings of a study conducted in low- and lower-middle-income countries and high-burden countries where the proportion of the pre-treatment loss to follow-up ranged from 4% to 38%.7 In two regions of Cameroon, the pre-treatment loss to follow-up rate was 16.7%.12 The Chongwe study conducted in the rural parts of Zambia reported a higher rate of pre-treatment loss to follow-up at 33%, with a high peak at over 80% in February of the rainy season;8 our study did not show any seasonal differences, possibly reflecting easier access to health facilities in the urban areas in Zambia regardless of the climate and road condition.
Implications
The findings of this study have some implications. The loss of bacteriologically confirmed patients with TB before treatment initiation represent an important failing in the provision of care by health facilities.13 In addition, untreated TB cases are key factors that contribute to the continued global TB epidemic,3–5 and high mortality has been reported in this group.14
Strengths
This study had several strengths. It is the first of its kind to be conducted in Lusaka, an urban setting and a high TB burden city. The sample size was large, with over 1,000 bacteriologically positive patients with presumptive TB at multiple health facilities enrolled in the study, representing about one-sixth of cases found in Lusaka; this makes the study results more representative than the previous study conducted in Zambia.8
Limitations
There are also several study limitations: as the study was conducted in a limited number of facilities, the findings do not represent pre-treatment loss to follow-up rate for the entire district of Lusaka or the country. There is a need to conduct a larger nation-wide study. Second, the patients who were excluded from the study for various reasons may have affected the findings. Third, reasons for not initiating treatment from the patient’s point of view could not be explored, as there was no direct contact with them.
In conclusion, health facilities should strive to account for every patient, particularly those who are bacteriologically confirmed, as this is the only way to curtail TB transmission in the community.
ACKNOWLEDGEMENTS
The authors thank all the staff members of Lusaka District Health Office and the National Tuberculosis and Leprosy Control Programme, the Ministry of Health, for their collaboration and support, without which this study could not have been completed.
Funding Statement
This study was financed through the routine activities of the Japan Anti-tuberculosis Association (Tokyo, Japan), funded by the Ministry of Foreign Affairs of Japan.
Footnotes
Conflicts of interest: none declared.
References
- 1.World Health Organization Global tuberculosis report, 2021. CC BY-NC-SA 3.0 IGO. Geneva, Switzerland: WHO; 2021. [Google Scholar]
- 2.National Tuberculosis and Leprosy Programme, Ministry of Health, Republic of Zambia Tuberculosis manual. 6th ed. Lusaka, Zambia: MoH; 2022. [Google Scholar]
- 3.Asres A, Jerene D, Deressa W. Delays to treatment initiation is associated with tuberculosis treatment outcomes among patients on directly observed treatment short course in Southwest Ethiopia: a follow-up study. BMC Pulm Med. 2018;18(1):64. doi: 10.1186/s12890-018-0628-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gebreegziabher SB, Bjune GA, Yimer SA. Total delay is associated with unfavorable treatment outcome among pulmonary tuberculosis patients in West Gojjam Zone, Northwest Ethiopia: a prospective cohort study. PLoS One. 2016;11(7):e0159579. doi: 10.1371/journal.pone.0159579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lee C-H, et al. Treatment delay and fatal outcomes of pulmonary tuberculosis in advanced age: a retrospective nationwide cohort study. BMC Infect Dis. 2017;17(1):449. doi: 10.1186/s12879-017-2554-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.World Health Organisation Joint Initiative “FIND. TREAT. ALL. #ENDTB”. Geneva, Switzerland: WHO; 2019. [Google Scholar]
- 7.MacPherson P, et al. Pre-treatment loss to follow-up in tuberculosis patients in low-and lower-middle-income countries and high-burden countries: a systematic review and meta-analysis. Bull World Health Organ. 2014;92(2):126–138. doi: 10.2471/BLT.13.124800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chilembo M, et al. Pre-treatment lost to follow-up tuberculosis patients, Chongwe, Zambia, 2017: a retrospective cohort study. Public Health Action. 2020;10(1):21–26. doi: 10.5588/pha.19.0059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Daka S, et al. Turnaround times of the sputum sample courier system at tuberculosis treatment centers in Lusaka, Zambia, 2021. Int J Mycobacteriol. 2022;11:103–107. doi: 10.4103/ijmy.ijmy_3_22. [DOI] [PubMed] [Google Scholar]
- 10.Pillai D, et al. Initial default among tuberculosis patients diagnosed in selected medical colleges of Puducherry: issues and possible interventions. Int J Med Sci Public Health. 2015;4(7):957–960. [Google Scholar]
- 11.Burger R, et al. Who is more likely to return for TB test results? A survey at three high-burden primary healthcare facilities in Cape Town, South Africa. Int J Infect Dis. 2021;113:259–267. doi: 10.1016/j.ijid.2021.10.015. [DOI] [PubMed] [Google Scholar]
- 12.Onyoh EF, Kuaban C, Lin HH. Pre-treatment loss to follow-up of pulmonary tuberculosis patients in two regions of Cameroon. Int J Tuberc Lung Dis. 2018;22(4):378–384. doi: 10.5588/ijtld.17.0676. [DOI] [PubMed] [Google Scholar]
- 13.Harries AD, et al. Registering initial defaulters and reporting on their treatment outcomes. Int J Tuberc Lung Dis. 2009;13:801–803. [PubMed] [Google Scholar]
- 14.Mugauri H, et al. Bacteriologically confirmed pulmonary tuberculosis patients: Loss to follow-up, death and delay before treatment initiation in Bulawayo, Zimbabwe from 2012–2016. Int J Infect Dis. 2018;76:6–13. doi: 10.1016/j.ijid.2018.07.012. [DOI] [PubMed] [Google Scholar]