Risk factors on length of stay among pulmonary tuberculosis patients: A systematic review and meta-analysis

Abstract

Background

Pulmonary Tuberculosis (PTB) remains a pressing public health concern. Long hospital stays for PTB patients can overburden both patients and healthcare systems.

Objective

To identify the key factors contributing to extended length of stay in PTB patients.

Information sources

Four electronic databases (PubMed, Scopus, Embase, and CINAHL) were systematically searched from inception to January 1, 2023.

Methods

The articles were screened and performed according to Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA). Inclusion criteria were PTB patients diagnosed by doctors and studies reporting factors affecting length of stay. Exclusion criteria were review articles, case study, conferences abstract, and proceedings. Study quality was assessed using the Newcastle-Ottawa Scale (NOS). A random-effects model was used to analyzed risk factors for length of stay. Heterogeneity was employed using I² and Q statistics. Forest plots displayed effect sizes (ES) and 95 % confidence intervals. STATA 14.2 was used for meta-analysis.

Results

A total of 1,190 studies were screened from reputable electronic databases, six studies comprised of 9,231 participants were included. Meta-analysis revealed that they are six risk factors associated with longer length of stay including; older age (OR 1.50, 95 % CI 1.07–2.09, p = 0.019), comorbidity (OR 1.44, 95 % CI 1.17–1.78, p = 0.001), HIV patient (OR 1.40, 95 % CI 1.16–1.69, p = 0.001), patients with ADR (OR 2.19, 95 % CI 1.47–3.26, p < 0.001), MDR TB (OR 3.16, 95 % CI 2.31–4.32, p < 0.001), and miliary TB (OR 1.37, 95 % CI 1.10–1.70, p = 0.004) with minimal heterogeneity [(I² = 34.2 %, p = 0.207), (I² = 43.1 %, p = 0.118), (I² = 0.0 %, p = 0.573), (I² = 0.0 %, p = 0.723), (I² = 0.0 %, p = 0.366), and (I² = 0.0 %, p = 0.753), respectively]. There was no evidence of publication bias according to Begg's and Egger's test.

Conclusions

In conclusion, six risk factors were identified as significantly associated with longer hospital stays in PTB patients: older age, comorbidities, HIV infection, ADR, MDR-TB, and miliary TB. These findings highlight the importance of targeted interventions for these high-risk groups to reduce length of stay and alleviate the burden on healthcare systems. The results are based on a meta-analysis of six studies with minimal heterogeneity, and no evidence of publication bias was found. Future research should focus on exploring additional factors influencing length of stay, particularly in diverse populations, and evaluating the effectiveness of interventions to shorten hospital stays. Additionally, studies examining the impact of healthcare infrastructure and resource allocation on length of stay could provide valuable insights for improving patient outcomes.

Registration

This study was registered with PROSPERO, CRD4203390615

1. Background

Pulmonary tuberculosis (PTB) is still one of the most common infectious disease related causes of mortality worldwide. The pathogen responsible for PTB, a deadly infection that mostly affects the lungs but can possibly spread to other organs, is Mycobacterium tuberculosis (MTB). Important risk factors include HIV infection, exposure to infection, and being born in an endemic country. The 6.2 million PTB cases that were reported worldwide in 2022 included 63 % of bacteriologically confirmed cases. This level was identical in 2021. Although the six WHO regions varied significantly, the Eastern Mediterranean Region had the lowest proportion (56 %) and the Region of the Americas had the highest incidence (79 %). There was also a great deal of variation among the countries. High-income countries had the highest levels of confirmation (median, 91 %) whereas low-income countries had the lowest (median, 71 %) in countries where the most sensitive diagnostic tests were widely available. Compared to 69 % (2.4/3.5 million) in 2021 and above the pre-pandemic level of 62 % (2.2/3.6 million) in 2019, 73 % (2.9/4.0 million) of individuals with bacteriologically proven PTB had rifampicin resistance testing in 2022 (World Health Organization., 2023). In 2021, TB caused 1.4 million deaths worldwide, ranking it as a top infectious disease killer after COVID-19. An estimated 10.6 million people fell ill with TB in 2021, highlighting its vast burden on global health systems (World Health Organization., 2023). PTB remains a critical public health issue globally due to its prevalence, transmission dynamics, economic and social impacts, and the challenges associated with treatment, especially in the context of drug-resistant strains. Effective PTB control requires comprehensive strategies, including timely diagnosis, adherence to treatment, robust healthcare systems, and global cooperation to reduce the burden of this disease.

Length of stay for TB patients can have significant effects on the psychological, social and economic well-being of patients and their families (Rucșineanu et al., 2018). These effects include: (1) physical health and cognitive function, where prolonged hospitalization increases the risk of contracting additional infections that can complicate TB treatment; extended bed rest leading to muscle weakness, reduced mobility, and overall physical deconditioning; altered cognitive function, whereby extended hospitalization, particularly in older adults, can lead to cognitive decline or delirium; declines in cognitive and physical functioning are the most common complications of hospitalization for older adults (Inouye et al., 2000). (2) psychological effects brought on by extended stays and prolonged separation from family and friends, such as feelings of isolation, anxiety, depression, and stress. especially older adults, hospitalization has been linked to psychosocial impairment. Patients in isolation had higher scores for depression and anxiety, and reported fear and loneliness (Abad et al., 2010). (3) Social and economic impacts, such as loss of income due to prolonged absence from work, leading to financial hardship; The longer duration of hospitalization for DR-TB, people with drug-resistant TB face a much greater economic loss (as a percentage of household income) than people with DS-TB (Rouzier et al., 2010; Tanimura et al., 2014). (4) burden on family, where families face economic and emotional strain due to medical expenses and the need for caregiving; and long-term hospitalization can cause economic burdens for patients and their families through loss of income and the cost of transportation and treatment (Abad et al., 2010). (5) the stigma of hospitalization for TB, which can exacerbate social discrimination, thereby affecting patients’ social relationships and community standing. Patients may feel stigmatized and without social support when they are cut off from their social networks (Rucșineanu et al., 2018).

Length of stay among PTB patients varies. A retrospective follow-up study showed that the duration of length of stay for intensive MDR-TB treatment ranged from 36 to 100 days, with the average stay being 76.7 days and the median being 62 days (Tamirat et al., 2020). Several factors were linked to longer hospital stays for patients with tuberculosis, including having the pulmonary form of the disease, being bedridden, and adverse drug reactions. Another study found that most PTB patients only needed to stay in the hospital for an average of 14 days, but the length of stay increased for patients with the more serious form of TB called miliary TB, for older patients, and for those admitted to certain hospitals. The length of stay for patients with PTB in hospitals is influenced by various factors, including clinical, demographic, and systemic variables. More specifically, these factors include: (1) severity of illness, where patients with severe PTB requiring complex treatment or critical care often have a longer length of stay; (2) TB-associated comorbidities that increase the length of stay; (3) non TB-associated comorbidities and underlying health conditions such as HIV, liver disease, anemia, malnutrition, and genitourinary tract diseases (Tonko et al., 2020); (4) treatment response, where patients who responded well to medication might be discharged sooner than those who experienced adverse reactions or slower recovery; (5) socioeconomic factors including access to healthcare facilities, social support networks, and financial resources; and (6) hospital protocols where different hospitals have varying criteria for discharge and treatment.

To date, there has been no study on which particular factors most strongly affect length of stay. Studies investigating factors influencing length of stay among PTB patients can provide valuable insights for healthcare practitioners, helping them to optimize patient management, improve outcomes, and allocate resources more effectively. Key findings from such studies typically identify clinical, demographic, and systemic factors that influence length of stay. Therefore, this study aimed to determine the length of stay and identify factors associated with extended length of stay during treatment.

2. Methods

This study followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines to ensure clear and transparent reporting, and is registered on PROSPERO, an international database for systematic review protocols, under number CRD4203390615. The study was conducted in accordance with ethical principles, and the protocol was approved by the University of Phayao Human Ethics committee under the reference number [HREC-UP-HSST 1.1/025/66].

2.1. Data sources and search strategy

In collecting relevant studies, we searched systematically several electronic databases including PubMed, Scopus, Embase, and CINAHL. The MeSH terms used were “tuberculosis,” “factor,” and “length of stay.” Systematic review and meta-analysis articles as well as publications reporting overlapping data were excluded. The references cited in the identified articles were also reviewed and judged for inclusion in case they were deemed relevant. The most recent search for studies included in this review was completed on January 1, 2023.

2.2. Study selection and data extraction

Regarding data extraction, three reviewers (WD, PJ, and WN) agreed on the criteria used to select studies for this review. The data was searched, screened, and extracted by five reviewers (WD, WN, PJ, UA and PS) and confirmed by either SS or UA. The search strings used and details are shown in Table S1. The authors discussed any uncertainties or differences in opinion to reach a final agreement on the approach.

Inclusion criteria were PTB patients diagnosed by doctors and studies reporting factors affecting length of stay. Exclusion criteria were review articles, case study, conferences abstract, and proceedings.

2.3. Quality assessment, risk of bias in included studies

The quality of the included non-randomized studies was assessed independently by two reviewers using the Newcastle-Ottawa Scale (NOS) (Higgins et al., 2003). The NOS awards up to 9 points, with studies scoring 7 or less considered of good quality. Discrepancies between reviewers were resolved through discussion, and if necessary, a third author was consulted (Higgins et al., 2003).

2.4. Data synthesis

The fully adjusted pooled overall odds ratios (ORs) with 95 % confidence intervals (CIs) from the primary studies were used to assess the risk factors associated with length of stay. If the primary studies reported effect estimates other than ORs (e.g., hazard ratios [HRs] or relative risks [RRs]), these estimates were converted to ORs for consistency using the following equations (Saokaew et al., 2021):

$$RR=\frac{OR}{\left(1-r\right)+\left(r*OR\right)}$$

and

$$HR=\frac{ln\left(1-\left(RR*r\right)\right)}{ln\left(1-r\right)}$$

where “r” is the event rate from all causes of the reference group.

A statistical analysis of the risk factors influencing length of stay was conducted with STATA 14.2. This study utilized a random-effects model to combine the results of several risk variables and their impact on durations of hospital stay. This approach was chosen to address heterogeneity found in the data. Forest plots were utilized to present the effect sizes (ES) from each research along their relevant 95 % confidence intervals and the overall estimated ES. The heterogeneity was assessed using I² and Q statistics (Egger et al., 1997). To categorize heterogeneity levels, recommended three categories: low (<25 %), moderate (25–75 %), and high (>75 %). We tried to find possible sources of heterogeneity in instances where it was found (Higgins et al., 2021). We carried out a subgroup analysis according to participant characteristics when suitable.

Subgroup analysis for each major outcome that has been pre-specified: participants' personality traits and our main predisposing factors will be investigated: model of analysis, residence, comorbidities including HIV, diabetes mellitus, and hypertension, and quality of the study (NOS). Publication bias was evaluated for each outcome using Egger's regression tests, Begg's test, and Funnel plot asymmetry tests (Higgins et al., 2021).

3. Results

3.1. Characteristics of included studies

A total of 1190 studies were screened from reputable electronic databases, six studies comprised of 9231 participants were included. Six studies were included, two articles described different study designs, three resulted in outcomes that were nonrelevant to the focus of this study, and three studies completed on unsuitable populations were excluded. This screening process resulted in six studies included in quantitative synthesis. Fig. 1 indicates a PRISMA flowchart that shows the method of selecting studies for review.

Fig. 1.

Selection flow diagram.

The six studies that met our inclusion criteria included research conducted in Ethiopia (Tamirat et al., 2020), Brazil (Gonçalves and Ferreira, 2013), Switzerland (Tonko et al., 2020), China (Zhou et al., 2019), Canada (Ronald et al., 2016), and Belgium (Cimpaye et al., 2019). Key features of each study have been analyzed including criteria of participants, sample size, participant demographics (age), intervention/exposure details, participant comorbidity details, NOS score. The important characteristics and outcomes of the included articles were collated and illustrated in Table 1.

Table 1.

Description of participants of included studies.

Author, Year	Participants	Criteria of participants		Age of participants (years old)		Sex of participants			Comorbidity of participants	Total NOS score
		Inclusion	Exclusion	Interventions	Comparators	Sex	Interventions	Comparators	Diseases
Tamirat et al. 2020	N = 432	Patients admitted and discharged from the selected hospitals during the initial phase of MDR-TB treatment were the study population.	Incomplete data, death during treatment, and transfer before completing the intensive phase. Data	< 24 (32.8 %) 25–34 (13.5 %) 35–44 (19 %) ≥ 45 (16.7 %)	–	–	–	–	HIV co-infection	5/9
Gonçalves and Ferreira, 2013	N = 306	Patients with TB, aged ≥15 years, both those previously diagnosed before admission and new cases identified during hospital stay, according to the diagnostic standard adopted by hospital medical services.	Patients who did not agree to be interviewed; those unable to answer questions due to mental confusion or inability to provide all information requested; and patients aged <18 years not accompanied by a legal guardian.	HIV positive ≤ 25 days (n = 89) 15–24 (11.2 %) 25–34 (50.5 %) 35–49 (31.5 %) ≥ 50 (6.7 %) HIV positive > 25 days (n = 63) 15–24 (14.3 %) 25–34 (49.2 %) 35–49 (28.6 %) ≥ 50 (7.9 %)	HIV negative ≤ 25 days (n = 107) 15–24 (16.8 %) 25–34 (16.8 %) 35–49 (26.2 %) ≥ 50 (40.2 %) HIV negative > 25 days (n = 47) 15–24 (8.5 %) 25–34 (14.9 %) 35–49 (38.3 %) ≥ 50 (38.3 %)	Male Female Male Female	HIV positive ≤ 25 days (n = 89) 57 (64.0 %) 32 (36.0 %) HIV positive > 25 days (n = 63) 39 (61.9 %) 24 (38.1 %)	HIV negative ≤ 25 days (n = 107) 58 (54.2 %) 49 (45.8 %) HIV negative > 25 days (n=47) 25 (53.2 %) 22 (46.8 %)	Hypertension and/or diabetes Hypertension and/or diabetes	8/9
Tonko et al., 2020	N = 6234	Patients in the database have a unique identifier that was used to track rehospitalizations.	–	< 14 days (n = 3154) < 25 (22.2) 25–39 (32.1) 40–64 (28.4) ≥ 65 (17.3)	> 14 days (n = 3080) < 25 (18.3) 25–39 (29.5) 40–64 (26.3) ≥ 65 (25.9)	Male Female	< 14 days (n = 3154) 1387 (44.0) 1767 (56.0)	> 14 days (n = 3080) 1162 (37.7) 1918 (62.3)	No comorbidity One comorbidity Several comorbidities	8/9
Zhou et al., 2019	N = 356	TB patients who gave informed consent for inclusion before participating in the study.	–	< 30 (19.7) 30–44 (13.5) 45–59 (27.8) ≥ 60 (39.0)	–	Male Female	270 (75.8 %) 86 (24.2 %)	–	–	5/9
Ronald et al., 2016	N = 1852	The participants were identified of all individuals with confirmed TB, and notified to Montreal Public Health. All persons diagnosed with microbiologically or clinically confirmed TB in Montreal.	–	0–19 (7.7) 20–34 (32.5) 35–64 (38.4) 65+ (21.1)	–	Male	992 (53.6 %)	–	Cancer Diabetes HIV positive Renal disease Liver disease Previous substance abuse Previous smoker	5/9
Cimpaye et al., 2019	N = 51	All patients who were admitted to respiratory isolation at Liège University Hospital for pulmonary tuberculosis during the period concerned	Patients whose diagnosis of tuberculosis was overturned after admission (n = 1), death preceding release from isolation (n = 3), release from isolation without doctor's approval (n = 2).	< 21 days (n = 29)	> 21 days (n = 22)	Male	21 (72,4 %)	14 (63.6 %)	HIV Diabetes Immunosuppressive treatment	5/9

These studies employed diverse methodologies across six distinct countries. One was a cross-sectional study (Zhou et al., 2019), three were retrospective studies (Tamirat et al., 2020, Ronald et al., 2016, Cimpaye et al., 2019), one was a cohort study (Gonçalves and Ferreira, 2013), and one was a case-control study (Tonko et al., 2020). The study sizes varied considerably, with the number of participants ranging between 306 and 6234, with specific participant counts as follows: 306 (Gonçalves and Ferreira, 2013), 356 (Zhou et al., 2019), 432 (Tamirat et al., 2020), 1852 (Ronald et al., 2016), and 6234 (Tonko et al., 2020), 51(Cimpaye et al., 2019).

3.2. Quality of included studies

The Newcastle-Ottawa Scale (NOS) was employed to assess the potential for bias within the studies included in this review. NOS revealed varying quality among the six included studies. Based on the combined evaluation of “selection,” “comparability,” and “ascertainment of exposure,” two studies achieved a high score of 8, while the remaining four received moderate scores of 5. Supportive details for these judgments are provided in supplementary file, Table S2.

3.3. Risk factor associated with longer length of stay

Meta-analysis revealed that they are six risk factors associated with longer length of stay including; older age (OR 1.50, 95 % CI 1.07–2.09, p = 0.019), comorbidity (OR 1.44, 95 % CI 1.17–1.78, p = 0.001), HIV patient (OR 1.40, 95 % CI 1.16–1.69, p = 0.001), patients with ADR (OR 2.19, 95 % CI 1.47–3.26, p < 0.001), MDR TB (OR 3.16, 95 % CI 2.31–4.32, p < 0.001), and miliary TB (OR 1.37, 95 % CI 1.10–1.70, p = 0.004) with minimal heterogeneity [(I² = 34.2 %, p = 0.207), (I² = 43.1 %, p = 0.118), (I² = 0.0 %, p = 0.573), (I² = 0.0 %, p = 0.723), (I² = 0.0 %, p = 0.366), and (I² = 0.0 %, p = 0.753), respectively]. (Fig. 2).

Fig. 2.

Forest plot showing odds ratio of length of stay and risk factors.

3.4. Publication bias and small-studies effect

There was no evidence of publication bias according to funnel plot. Additionally, Begg's and Egger's tests did not indicate evidence of a small-studies effect; the results of Begg's test (p = 0.828) and Egger's test (0.823) indicated no significant differences in any age groups or outcomes. (Supplementary file, Fig. S1).

3.5. Sensitivity and subgroup analysis

Sensitivity and subgroup analyses were conducted to assess the robustness of effect estimates (Table 2). Switching from fixed-effects to random-effects models showed no significant changes in pooled ORs across outcomes, with results consistent with the main analysis. Key findings include: (1) Fixed- and random-effects models produced consistent results; (2) Subgroup analyses revealed that comorbidities and miliary TB contribute to prolonged length of stay in developing countries, with ADRs remaining significant; (3) NOS scores indicated that HIV, and comorbidities are associated with longer length of stay, even in studies with scores below 7.

Table 2.

Sensitivity and subgroup analyses.

Characteristic	Older age			HIV infection			ADRs			MDR-TB			Comorbidities			Miliary TB
	OR (95 % CI)	Heterogeneity		OR (95 % CI)	Heterogeneity		OR (95 % CI)	Heterogeneity		OR (95 % CI)	Heterogeneity		OR (95 % CI)	Heterogeneity		OR (95 % CI)	Heterogeneity
		I2(%)	p		I2(%)	p		I2(%)	p		I2(%)	p		I2(%)	p		I2(%)	p
Models
Fixed effect model	1.72 (1.49, 1.99)	34.2	0.207	1.40 (1.16, 1.69)	0.0	0.573	2.19 (1.47, 3.26)	0.0	0.723	3.16 (2.31, 4.32)	0.0	0.366	1.42 (1.24, 1.64)	43.1	0.118	1.37 (1.10, 1.70)	0.0	0.753
Random effects model	1.50 (1.07, 2.09)	34.2	0.207	1.40 (1.16, 1.69)	0.0	0.573	2.19 (1.47, 3.26)	0.0	0.723	3.16 (2.31, 4.32)	0.0	0.366	1.44 (1.17, 1.78)	43.1	0.118	1.37 (1.10, 1.70)	0.0	0.753
Country*
Developed	1.81 (1.55, 2.11)	–	–	1.43 (1.16, 1.76)	–	–	N/A	N/A	N/A	3.16 (2.31, 4.32)	0.0	0.366	1.43 (1.11, 1.86)	63.3	0.043	1.37 (1.10, 1.70)	0.0	0.753
Developing	1.13 (0.72, 1.77)	0.0	0.681	1.23 (0.76, 1.99)	–	–	2.19 (1.47, 3.26)	0.0	0.723	N/A	N/A	N/A	1.50 (0.91, 2.47)	0.0	0.444	N/A	N/A	N/A
Quality NOS
NOS ≥ 7	1.80 (1.55, 2.10)	0.0	0.574	N/A	N/A	N/A	2.52 (1.05, 6.05)	–	–	3.95 (2.22, 7.02)	–	–	1.82 (0.90, 3.70)	–	–	1.40 (1.08, 1.82)	–	–
NOS < 7	1.12 (0.70, 1.80)	0.0	0.382	1.40 (1.16, 1.69)	0.0	0.573	2.11 (1.35, 3.30)	–	–	2.88 (1.99, 4.17)	–	–	1.42 (1.13, 1.78)	51.9	0.081	1.30 (0.89, 1.90)	–	–

^⁎According to UNIDO, 2024 (https://stat.unido.org/portal/storage/file/publications/country-classif-report-2024.pdf)

4. Discussion

This study synthesizes research on factors affecting length of stay for patients with PTB, analyzing data from four databases: PubMed, SCOPUS, Embase, and CINAHL. After screening 1190 papers, six relevant studies were selected. Sensitivity and subgroup analyses showed no significant change in pooled odds ratios when switching from fixed-effects to random effects models. The analysis identified six factors influencing length of stay in PTB patients: (1) older age, (2) HIV infection, (3) adverse drug reaction (ADR), (4) multidrug-resistant (MDR) TB, (5) comorbidities, and (6) miliary TB, which are discussed in detail in subsequent paragraphs.

Age significantly impacts the length of stay for PTB patients, as older individuals typically have weakened immune systems, making it harder to fight the infection and often resulting in longer hospitalizations (Caraux-Paz et al., 2021). Their illness may be prolonged, especially with limited access to healthcare in remote areas. (Weiangkham et al., 2014). Economic and social factors, such as housing conditions and caregiver availability, complicate recovery, increasing hospital occupancy and costs (Hendy et al., 2012). Modifiable risks like caregiver stress and nursing home placement also influence length of stay (Toh et al., 2017). The effect of age on length of stay varies based on individual health and social support.

HIV infection significantly impacts length of stay for PTB patients, as TB is the leading cause of death among those living with HIV/AIDS (Gupta et al., 2015). The co-occurrence of HIV and TB is common, as HIV weakens the immune system, increasing susceptibility to TB (Balasubramaniam et al., 2019). For example, women with HIV have hospital stays that are 30 % longer after childbirth compared to those without HIV, and the length of stay for women with TB or both TB and HIV can be more than twice as long as for healthy mothers (Falana et al., 2018). HIV coinfection can lead to advanced TB disease, necessitating prolonged monitoring and recovery times (Hasan et al., 2010). Additionally, patients with TB may face complications like pneumonia during extended hospitalizations, further increasing their length of stay, particularly for those with symptomatic HIV (Rowell-Cunsolo et al., 2018).

Adverse drug reactions (ADRs) lead to significantly longer length of stay for PTB patients. First-line TB drugs like INH, RIF, and PZA, while effective, can cause liver damage and various side effects such as gastrointestinal issues, allergies, joint pain, and nervous system disorders (Lv et al., 2013; Tostmann et al., 2008). These unpleasant side effects may hinder patient adherence to treatment, which is particularly concerning given the rise of drug-resistant TB and the scarcity of effective medications (Yang et al., 2019). The most common ADRs include cutaneous adverse drug reactions (21.0 %), drug-induced hepatitis (7.1 %), and gastrointestinal disturbances (4.8 %), with pyrazinamide being the most frequent culprit. A notable portion of patients (15.7 %) required treatment adjustments due to ADRs, which can lead to a doubling of hospital stays, extending them by up to 20 days (Davies et al., 2009; Fei et al., 2018). Managing ADRs effectively is essential for improving patient outcomes, alleviating healthcare resource strain, and reducing hospital LOS.

Multidrug-resistant tuberculosis (MDR-TB) is a more resilient form of TB resistant to isoniazid and rifampin, significantly prolonging the length of stay for PTB patients (World Health Organization., 2023). In Bhutan, the misuse of medications has contributed to the rise of MDR-TB (Tenzin et al., 2020). Treatment plans for drug-resistant TB are based on resistance studies across populations, but the lack of susceptibility tests for all anti-TB drugs means individualized regimens are rare, leading to standardized treatments as seen in Ethiopia (Kebede et al., 2014). Treatment for drug-resistant TB lasts 18 to 24 months, but the cure rate is only about 50 % (Oladimeji et al., 2016). Addressing drug resistance is crucial for improving patient outcomes, alleviating healthcare system burdens, and shortening LOS. Comorbidities significantly increase the length of stay for PTB patients, especially when multiple conditions are present. In the context of PTB patients, comorbidities such as diabetes, cardiovascular disease, and respiratory conditions can complicate TB management and increase susceptibility to complications, leading to longer length of stay For instance, the interplay between TB and diabetes mellitus can result in prolonged treatment and higher recurrence rates (Khattak et al., 2024). People with both TB and DM are more likely to experience problems tolerating TB medications, including increased side effects, toxicity from the medications, and interactions between the TB and DM medications (Abbas et al., 2022). Similarly, chronic obstructive pulmonary disease (COPD) can worsen TB progression, making it more severe and prolonged (Jakimova et al., 2019). Co-infections with other diseases pose additional challenges, as TB medications can be hepatotoxic and interact adversely with treatments for other conditions, reducing recovery chances (Dzinamarira et al., 2022). Hepatitis B co-infection can exacerbate TB symptoms, leading to poorer outcomes, and patients with hepatitis C and B often experience significant interruptions in TB treatment (Chua et al., 2018). Overall, comorbidities necessitate increased medical attention, contributing to extended treatment times due to drug interactions and complications.

Miliary tuberculosis typically results in longer length of stay due to its serious nature, as the bacteria spread through the bloodstream and form tiny nodules in various organs (Sharma and Mohan, 2024). This form of TB is associated with poor prognosis (Wakamatsu et al., 2018) and complicates management, often requiring intensive and prolonged treatment regimens. Miliary TB has high morbidity and mortality rates, with treatment delays being a significant factor in mortality (Sharma and Mohan, 2024). The infection's systemic nature necessitates careful monitoring and can lead to increased length of stay due to the complexities involved (Tonko et al., 2020). This highlights the need for specialized care and additional resources to effectively manage patients with miliary TB.

5. Strengths and limitations

We adhere to the strict requirements for conducting a thorough systematic review. In addition, our study included eliminating language barriers, using an existing protocol created to answer a research question, identifying pertinent studies, rigorously evaluating the methodological quality of the included studies. Furthermore, there is some generalizability in our findings. We integrated research including various patient demographics, nations, medical environments, ages, and care systems. Our findings can therefore be used in the majority of healthcare environments across the globe. Additionally, we advise nurses and other medical professionals to educate patients and/or other healthcare professionals on how to prevent and manage PTB-related length of stay. Based on our research, these variables might indicate the distinctiveness of PTB's LOS. However, several limitations of our review should be considered. First, the populations included in the studies came from a range of countries and healthcare settings, which could lead to significant variability in treatment standards and care systems. Despite these potential differences, the primary factors contributing to prolonged length of stay in individuals with PTB remained consistent. Our sensitivity analysis confirmed that the predictors of length of stay were stable across studies, which suggests that the findings may be generalizable to PTB patients in diverse setting. Second, the type of comorbidity was a crucial factor that may have influenced our results. Regardless of the specific comorbidities, the sensitivity analysis revealed a consistent trend towards longer length of stay in PTB patients with comorbid conditions, highlighting the importance of these factors in the overall length of stay outcomes. Third, while we conducted a thorough review of several major electronic databases, including PubMed, Scopus, Embase, and CINAHL, it is possible that studies published in local databases or in countries not covered by these resources were not included in our analysis. However, our evaluation of publication bias, using the funnel plot, Begg's test, and Egger's test, did not reveal any significant evidence of bias, which suggests that the results of our review are not substantially influenced by unpublished studies or those from underrepresented regions. Fourth, different studies may utilize slightly varying definitions of length of stay, which could potentially lead to discrepancies in the results and distort the overall effects of factors related to length of stay. These variations in how length of stay is defined might affect the comparability of findings across studies. However, in our sensitivity analysis using average length of stay for PTB patients (<20 days vs ≥21 days), we found that the effect estimates remained consistent (OR 1.68, 95 % CI 1.41–1.96 versus OR 1.31, 95 % CI 0.96–1.66). This suggests that these differences in definition did not significantly impact the overall conclusions. This stability of the results reinforces the reliability of the observed relationships between the factors and length of stay in the context of PTB patients.

6. Conclusions and implications

This study identifies key factors influencing the length of stay for pulmonary tuberculosis (PTB) patients, including older age, comorbidities, HIV infection, adverse drug reactions (ADRs), multidrug-resistant TB (MDR-TB), and miliary TB. These findings underscore the importance of early diagnosis and the ability to predict length of stay to prevent unnecessary hospitalizations, enhance patient care, and optimize resource allocation. Despite healthcare system variations, predictors of length of stay were generally consistent, highlighting their generalizability across different settings.

To reduce length of stay and improve outcomes, clinicians should prioritize early management of comorbidities and other risk factors. Policymakers should work toward standardizing PTB treatment protocols and improving access to resources, particularly for managing comorbid conditions. Additionally, public health specialists and healthcare professionals can use these insights to guide evidence-based interventions, improve TB management, and reduce healthcare costs.

Further research is needed to explore the mechanisms linking comorbidities to prolonged length of stay and to validate these findings in diverse populations, ensuring tailored interventions for effective TB care worldwide.

Data availability

Data will be available upon request.

Funding sources

This work was partially funded by University of Phayao under Unit of Excellence on Clinical Outcomes Research and Integration (UNICORN). The funding source had no role in any of the decisions taken in planning and conducting the project or publishing the results.

CRediT authorship contribution statement

Dao Weiangkham: Writing – review & editing, Writing – original draft, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Adinat Umnuaypornlert: Writing – review & editing, Validation, Methodology, Formal analysis, Conceptualization. Surasak Saokaew: Writing – review & editing, Validation, Supervision, Methodology, Conceptualization. Neeranuch Wongcharoen: Validation, Software, Resources, Methodology, Formal analysis, Data curation. Samrerng Prommongkol: Validation, Software, Resources, Investigation, Data curation. Jutamas Ponmark: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.