Tuberculosis diagnostic delay and its correlates among patients hospitalized during the COVID-19 pandemic period at a tertiary public hospital in Nairobi, Kenya

INTRODUCTION

TB though an ancient disease, still remains a significant public health problem and the sole cause of global mortality from a single infectious agent.^1,2 The disease is transmitted by Mycobacterium tuberculosis bacteria found in tiny droplets generated by infected individuals through laughing, coughing and sneezing.^3,4 Globally, approximately one quarter of the population is infected with M. tuberculosis.⁵ Africa accounts for 24% of global incident cases (2.5 million) whereas sub-Saharan Africa contributes 29% of all the TB cases worldwide.⁵ Based on the 2024 WHO global TB report, almost 10.8 million people developed TB in 2023, an increase from 10.7 million in 2022 and much higher than 10.4 million in 2021 and 10.1 million in 2020 hence led to the drastic negative effects in the global TB control progress.²

This upward trend reflects the ongoing after-effects of TB services disruptions during the COVID-19 pandemic period.⁵ Out of the 140,000 people who developed TB in Kenya in 2020, only 72,943 were diagnosed and notified meaning almost 50% of TB cases were either undiagnosed or unnotified. Nairobi County contributed just 9,804 (13%) (down from 15 %) to the national notifications.⁴ Meanwhile, KNH that normally accounts for about 10% of the Nairobi County case notification nosedived to just 618 (3.4%) cases.^6,7 Early TB diagnosis enhances effective disease prevention and control both in the community and in the health care facilities.⁸ Longer duration between symptom development and diagnosis increases TB-related morbidity and mortality.³ In Kenya, TB cases identification is mainly passive where after symptoms development, patients visit health care facilities 3 leading to late diagnosis.⁹

The emergence of (COVID-19) in 2020, reversed the global TB control progress made previously leading to a massive decrease in newly diagnosed and notified TB cases and consequently pushing the global TB targets off track.¹⁰ During the COVID-19 pandemic period, challenges such as unreliable transport as well as government movement restrictions could have contributed to delay in TB diagnosis.¹¹ Therefore, this study determined TB diagnostic delay and its correlates among patients hospitalized during the COVID-19 pandemic period at KNH in Nairobi, Kenya to contribute to the country's national policy in effective TB patient management in case of future public health emergencies.

METHODS

Study design, site, participants and sampling

This was a cross-sectional hospital-based mixed parallel study that involved TB records review and in-depth telephone interviews of TB patients after being discharged from hospital. Quantitative and qualitative data were simultaneously collected and analyzed separately. The study was conducted at KNH-a level six 1800 bed - capacity hospital in Nairobi County. The hospital has eight TB in-patient units that handle hospitalized TB cases, two TB laboratories (one for GeneXpert and the other for sputum smear microscopy tests) and a Radiology department for radiologic TB diagnostic tests.

The study participants were newly diagnosed adult TB patients hospitalized from January 2020 to December 2021. Purposive sampling was used to select the study site; census was used to select records of all adult TB patients. Incomplete records, <18 years, absconded patients and deaths were excluded. Qualitatively, simple random sampling was used to select IDI interviewees.

Data collection

Quantitatively, a refined Microsoft Excel 2016 spreadsheet was used to extract data from the adult inpatients’ records including demographics, clinical characteristics, TB tests done, results turn - around time, date of admission and date of TB diagnosis. Qualitatively in-depth telephone interviews were conducted on 36 patients discharged home, randomly selected, orally consented to participate in the study and audio-recorded. The interviews included awareness of TB signs and symptoms, health facilities visited, barriers on the pathway to final TB diagnosis.

Data management and analysis

Quantitative data was coded, entered into excel spreadsheet and edited for errors before analysis using Stata software, version 15.0 as per the study objectives. Data summaries and generated frequencies on categorical variables were conducted using descriptive statistics and presented in text, tables and figures. Inferential analysis at 95% CI was used to check for possible association between independent and dependent variables. Diagnostic delay was defined as the duration between admission and diagnosis which was 5 days.

Only characteristics potentially associated with diagnostic delay (>5 days) were evaluated using logistic regression in both bivariate and multivariate. All variables with a p value≤0.25 in bivariate analysis were subjected to multivariate analysis to establish the factors that were independently associated with TB diagnostic delay. All variables with a p<0.05 were considered statistically significant. Qualitatively, in-depth interviews were transcribed and analyzed thematically.

A six-step process was used in analyzing the data (familiarization, coding, themes generation, themes review, defining and naming of the themes and documenting the data analysis. First the scripts were read and re-read to familiarize with initial codes and identify the relevant features. Codes were then collated into potential themes which were reviewed further before final themes were generated and identified in form of topics, ideas and patterns or repeated meaning.

A matrix on emerging themes was created against the patients’ responses by use of N-Vivo software version 12. Data was analyzed by two researchers independently and interpretations of themes were shared and discussed to agree upon key themes that had emerged from the data to improve inter-rater reliability.

Ethical approval

Ethical approval was sought from the KNH/UoN Ethics and Research Committee (P657/08/2022), clearance from NACOSTI (P-23-27049) and permission from JKUAT post graduate school before the study was commenced. Participation in this study was voluntary and there were no incentives for participants. Thorough explanation on the aim and study purpose was done to the study participants before obtaining informed oral consent from them. Confidentiality was assured on all the information provided by restricting persons accessing the data.

RESULTS

Characteristics of patients hospitalized at KNH during the COVID -19 pandemic period

A total of 839 patients hospitalized during the study period were selected into the study of whom 258 (24.6%) were in 2020 and 581 (55.4%) in 2021.

About 494 (58.9%) were males, 580 (69.2%) aged between 18-44 years, 535 (63.8%) had normal BMI, 672 (80.1%), extra pulmonary TB (EPTB) of whom 327 (48.7%) had central nervous system TB (CNS TB, 394 (47%) were HIV positive and all of them were on cotrimoxazole preventive therapy and HAART) (Table 1).

Investigations used in final TB diagnosis during the COVID -19 pandemic period

A total of 796 (94.9%) patients were diagnosed using radiologic tests (X-ray, U/S, CT scan or MRI). Out of 36 patients eligible for GeneXpert test in 2020, only 14 (38.9%) were diagnosed (Table 2).

Factors associated with TB diagnostic delay during the pandemic period

A total of 563 (67.1%) delayed. EPTB was significantly associated with TB diagnostic delay aOR=4.108 (95% CI 2.782–6.067), p–0.000. Conversely, CNS TB was significantly protective against TB diagnostic delay aOR=0.243 (95% CI 0.165–0.359), p–0.000 compared to other forms of EPTB. Nonuse of radiologic tests aOR=6.521 (95% CI 1.519–27.996), p–0.012 and use of GeneXpert test aOR=6.306 (95% CI 3.763–10.568), p–0.000 were significantly associated with TB diagnostic delay (Table 3).

Barriers on the patients' pathway to final TB diagnosis

Personal/patient (restricted movement, fear testing positive for COVI-19, fear of forceful isolation if tested positive for COVID-19 and low awareness on TB symptoms) were the identified barriers to final diagnosis. Private (multiple visits to the same facility) and public facilities (conversion of public facilities to COVID-19 centers and long test result TAT) contributed to TB diagnostic delay (Table 4).

*Radiologic tests – X-ray, Ultra sound, CT scan and MRI

*others–pleural effusion, pericarditis, disseminated, peritonitis, milliary, bone, spine, eye, tongue, skin, testes, vulva, ovaries, **Radiologic tests–X-ray, ultrasound, CT scan and MRI

DISCUSSION

Majority of patients in this study were aged between 18–44 years, which concurs with findings from a similar study in Kenya as well as the 2016 national prevalence survey 15 where TB was found to be common between 18 to 50 years.¹⁴ Despite documented evidence that low BMI is associated with TB, most patients in this study had a normal BMI. This could be due to minimal impact of EPTB on appetite and metabolism.^14,16 Having >2/3 of patients hospitalized during the pandemic period delay in getting a TB diagnosis concurs with WHO predictions that due to disruptions of TB services by the COVID-19 pandemic leading to increased diagnostic delay, underdiagnosis and TB-related mortality.² Furthermore, government movement restrictions and conversion of lower public facilities to COVID-19 centers forced patients to either visit private or tertiary facilities thus worsening the diagnostic delay.¹⁷ In this study, EPTB was significantly associated with TB diagnostic delay probably due to inaccessibility and the paucibacillary of the affected sites as well as lack of universal standardized and efficient EPTB sample processing techniques hence delaying the diagnostic process.^18–20 Inversely, CNS TB was protective against delayed diagnosis contrary to similar studies from other high TB-burden countries (China, India and Indonesia) where patients with CNS TB were more likely to have a delayed diagnosis due to hesitancy of clinicians to perform TBM diagnostic procedures like lumbar punctures and low sensitivity of these diagnostic tests on EPTB.^21,22 The difference in these findings could be due to diverse study designs, TB control models, regional and facility-specific diagnostic guidelines and probably clinicians in our study site relying on clinical rather than microbiological confirmatory diagnosis hence reducing diagnostic delay. Despite being the gold standard in TB diagnosis due to its sensitivity, quick and simultaneous rifampicin resistance detection, in this study, GeneXpert test was significantly associated with TB diagnostic delay.²¹ This is also contrary to other similar studies which revealed that GeneXpert test exponentially improved. TB and rifampicin resistance detection thereby promoting TB control and improving patient outcomes.^17,23,24 The delay in this study could be explained by the fact that laboratory staff handling sputum samples in the study site feared contracting COVID-19 thereby delaying the sample processing and release of results.⁷

On their way to final diagnosis, a substantial number of patients in this study made multiple visits to private facilities without being diagnosed while others were treated severally in the same private facilities for non-TB conditions without referral to other facilities. Previous studies have documented reluctance of private practitioners to refer TB patients as a challenge in TB care and management.^9,14 Therefore, effective use of standardized diagnostic protocols and procedures can enhance early TB diagnosis. Long test result waiting time and conversion of public facilities to COVID-19 centers were also some of the barriers identified in public health facilities during the COVID-19 period.

This concurs with the 2020 Kenya national TB program report which indicated that conversion of some health care facilities to COVID-19 solation centres impacted negatively on diagnosis and follow up of patients in those facilities.⁷ This led to staff shortage 3 relocation of staff to COVID-19 centers culminating to long waiting times in health facilities.²⁵ In this regard, ensuring availability of adequate resources, minimal disruption of existing services as well as effective collaboration between the public and the private sector in TB diagnosis and management could prevent any possible future disruptions and improve TB service delivery while minimizing delays in case of future pandemics.¹⁴

Limitations

Using retrospective data collection in the quantitative arm could have higher likelihood of incomplete or missing data. Additionally, in a cross-sectional study, it becomes difficult to make causal inferences besides providing contrary findings had another timeframe been used.

CONCLUSION

As much as hospitalization is supposed to favor prompt TB diagnosis, reduce HCF disease transmission risks, improve patient outcomes and HC system efficiency, TB diagnostic delay worsened among patients hospitalized in KNH during the COVID-19 period. Moreover, EPTB, nonuser of diagnostic radiologic and use of GeneXpert tests were significantly associated with delay. Furthermore, personal, private and lower public facilities’ barriers contributed to final diagnosis. The health authorities need to promote effective public private partnerships in TB management & control during future pandemics. Further studies should be conducted to establish why high rates of EPTB in this study site.

Funding: No funding sources

Conflict of interest: None declared

Ethical approval: The study was approved by the Institutional Ethics Committee

REFERENCES

1. 1. MOH. Kenya Tuberculosis Annual Report. 2021.
2. 2. WHO 2020. Global tuberculosis report. Geveva: World Health Organization. 2020.
3. 3. Masini E, Hanson C, Ogoro J, Brown J, Ngari F, Mingkwan P, et al. Using Patient-Pathway Analysis to Inform a Differentiated Program Response to Tuberculosis: The Case of Kenya. J Infect Dis. 2017;216(4):714–23.
4. 4. Patterson B, Wood R. Is cough really necessary for TB transmission. Tuberculosis. 2019;117:31–5.
5. 5. WHO. Global Tuberculosis Report 2024. Geneva: World Health Organization. 2024.
6. 6. Makori L, Gichana H, Oyugi E, Nyale G, Ransom J. Tuberculosis in an urban hospital setting: Descriptive epidemiology among patients at Kenyatta National Hospital TB clinic, Nairobi, Kenya. Int J Africa Nurs Sci. 2021;15:100308.
7. 7. Enos M, Sitienei J, Ong’ang’o J, Mungai B, Kamene M. Kenya tuberculosis prevalence survey 2016: challenges and opportunities of ending TB in Kenya. PloS one. 2018;13(12):209098.
8. 8. Derseh D, Moges F, Tessema B. Smear positive pulmonary tuberculosis and associated risk factors among tuberculosis suspects attending spiritual holy water sites in Northwest Ethiopia. BMC Infect Dis. 2017;17(1):100.
9. 9. Kunjok DM, Mwangi JG, Mambo S, Wanyoike S. Assessment of delayed tuberculosis diagnosis preceding diagnostic confirmation among tuberculosis patients attending isiolo county level four hospital, Kenya. Pan Afr Med J. 2021;38:743.
10. 10. WHO. Tuberculosis report: World Health Organization. January. Geneva: World Health Organization. 2022.
11. 11. Hopman J, Allegranzi B, Mehtar S. Managing COVID-19 in Low- and Middle-Income Countries. JAMA. 2020;323(16):1549.
12. 12. Dawadi S, Shrestha S, Giri RA. Mixed-Methods Research: A Discussion on its Types, Challenges, and Criticisms. JPSE. 2021;2(2):25–36.
13. 13. Dworkin SL. Sample size policy for qualitative studies using in-depth interviews. Archives of Sexual Behavior. 2012;41(6):1319–20.
14. 14. Kiarie H, Temmerman M, Nyamai M, Liku N, Thuo W, Oramisi V, et al. The COVID-19 pandemic and disruptions to essential health services in Kenya: a retrospective time-series analysis. The Lancet Global Health. 2022;10(9):1257–67.
15. 15. Writing Committee of the Annual Report on Cardiovascular Health and Diseases in China. Interpretation of the annual report on cardiovascular health and diseases in China 2020. Cardiol Disc. 2022;2(04):269-85.
16. 16. Casha AR, Scarci M. The link between tuberculosis and body mass index. J Thorac Dis. 2017;9(3):301–3.
17. 17. Khan A, Khan N, Singh R. Tuberculosis diagnosis versus GeneXpert® MTB/RIF formats. Bioanalysis. 2024;16(16):843–8.
18. 18. Pradipta IS, Idrus LR, Probandari A, Puspitasari IM, Santoso P, Alffenaar JWC, et al. Barriers to Optimal Tuberculosis Treatment Services at Community Health Centers: A Qualitative Study From a High Prevalent Tuberculosis Country. Front Pharmacol. 2022;13:857783.
19. 19. Saktiawati AMI, Putera DD, Setyawan A, Mahendradhata Y, Van Der Werf TS. Diagnosis of tuberculosis through breath test: A systematic review. eBioMedicine. 2019;46:202–14.
20. 20. Solovic I, Jonsson J, Korzeniewska-Kosela M, Chiotan DI, Pace-Asciak A, Slump E, et al. Challenges in diagnosing extrapulmonary tuberculosis in the European Union, 2011. Eurosurveillance. 2013;18(12):1–9.
21. 21. Purohit MR, Purohit R, Mustafa T. Patient Health Seeking and Diagnostic Delay in Extrapulmonary Tuberculosis: A Hospital Based Study from Central India. Tubercul Res Treat. 2019;2019:1–8.
22. 22. Imran D, Satiti S, Sugianto P, Estiasari R, Maharani K, Pangeran D, et al. Barriers to diagnosis and management of CNS infections in Indonesia. Neurology. 2019;92(2):104–6.
23. 23. Nweke R, Eleazar C, Nweke B, Anaele E, Emenuga V. GeneXpert MTB/RIF Assay of Pulmonary Tuberculosis and HIV Co-Infection at a Central TB Referral Centre in Anambra State, Nigeria. Microbe Infect Dis. 2025;6(3):1084-93.
24. 24. Kabir S, Parash MTH, Emran NA, Hossain ABMT, Shimmi SC. Diagnostic challenges and Gene-Xpert utility in detecting Mycobacterium tuberculosis among suspected cases of Pulmonary tuberculosis. Quinn F, editor. PLoS ONE. 2021;16(5):251858.
25. 25. Suseela RP, Shannawaz M. Engaging the Private Health Service Delivery Sector for TB Care in India—Miles to Go. Tropical Med. 2023;8(5):265.