Background
Severe community-acquired pneumonia related treatment failure is persistence of features of severe pneumonia after initiation of antimicrobial therapy or a worsening clinical condition within 48–72 hours of the commencement of the antibiotics [1,2]. Community-acquired pneumonia is a bacterial, viral, and fungal lower respiratory infection that is not acquired at health facilities and can be manifested by persistent dry cough, elevated body temperature (more than 38 °C), chills, fatigue, difficulty breathing, rigours, and pleurisy chest pain [3]. It is the most frequent reason for hospitalization in children, and it costs a large amount of health system resources [4]. Severe community-acquired pneumonia (SCAP) is a life-threatening disease (the severe form of community-acquired pneumonia) that requires treatment in the intensive care unit and results in a higher mortality rate [5]. The prevalence of severe community-acquired pneumonia (SCAP) ranges from 5–35% in all CAP cases [6], and its fatality accounts for about 20–50% [7]. The rate of treatment failure among CAP patients is 11% [8,9]. The CAP is the most common cause of emergency care; about 75% of the CAP patients need emergency treatment [10].
Treatment failure in SCAP is a common condition in which an inadequate response to the antibiotic treatment can lead to the worsening of the disease and cause death [2]. Once treatment fails, the pathogen is disseminated out of the lung and causes empyema, meningitis, and endocarditis, which finally becomes a systemic disease and ends in death [2,11]. The timely risk classification and prediction of treatment failure for SCAP patients is crucial to rescuing therapeutic strategies and reversing the unfavorable outcome associated with SCAP [12].
Treatment failure causes prolonged hospitalization, reduces the quality of care, and places an places an economic burden on the health care system [13–17]. To reverse these impacts, identification of those children with a high risk of treatment failure is very important to close follow-up on their treatment response and early identification.
About 80% of the treatment failure of SCAP is because of the systemic inflammatory response [12]. The previous risk factor studies also showed that the most common causes of treatment failure in SCAP were host factors, appropriate treatment regimen [18], age, renal and cardiac comorbidity [19], infancy, malnutrition, severity of anaemia, rickets, lack of immunization, and hypoxia at baseline [20]. Even though the treatment failure associated with SCAP is the most devastating public health problem in Ethiopia, there is no sufficient studies to derivate and validate the model to predict the treatment failure among children with SCAP. Therefore, the current study aimed to assess the derivation and validation of a model to predict treatment failure among children with SCAP who are admitted to the Debre Tabor specialized comprehensive hospital intensive care unit.
Method
Study design, setting and period
A retrospective cohort study was used among children with SCAP who were admitted to the Debre Tabor specialized hospital intensive care unit. Debre Tabor (north central Ethiopia) is the administrative town of the south Gondar zone. It is located 103 km from the capital city of the Amhara region, namely Bahir Dar city, and 666 km from Addis Ababa (the capital city of Ethiopia). The geographic positioning system coordinates of Debre Tabor town are located at latitude (11.8500) and longitude (38.0167), and the average temperature ranges from 50°F to 76°F. Debre Tabor has a total of three health centers, one specialized comprehensive hospital, and six health posts. The data were conducted from November 27, 2022, to December 31, 2022.
Population (domain) and data sources
The source population was all under five children with SCAP admitted and treated at any time in the Debre Tabor comprehensive specialized hospital intensive care unit, and under five children with SCAP during the study period were included in the study. We used the data from the previous research [21].
Outcome of interest
Treatment failure (Yes, No).
Independent variables
Socio-demographic characteristics.
Age of the child and sex of the child.
Baseline disorder/status.
Pulmonary tuberculosis, meningitis, diarrhoea, HIV status, dehydration, pleural effusion, anaemia, hypoglycemia, measles, chest indrowing, previous history of SCAP, weight for age, status at admission, pertussis, and pulse rate (all these independent variables were determined at the initial presentation).
Operational definition
Treatment failure.
Was measured as persistence of features of severe pneumonia (cough and/or difficult breathing, with or without fever, fast breathing, or a lower chest wall in drawing where their chest moves in or retracts during inhalation) after initiation of antimicrobial therapy or a worsening clinical condition within 48–72 hours of the commencement of the antibiotics [1,2]. The outcome was measured or determined during their hospital stay.
The SCAP was defined as a CAP associated with the presence of one major or two or more minor criteria. The major criteria included a need for mechanical ventilation and septic shock. The minor criteria included systolic BP ≤ 90 mmHg, bilateral pneumonia or multilobar pneumonia, and PaO2/FIO2 ≤ 250 [22].
Comorbidity.
A disease condition at admission in addition to SCAP like heart disease, childhood asthma, retroviral infection, tuberculosis, acute gastroenteritis, pertussis, anaemia, meningitis, measles, bronchitis, heart disease, urinary tract infection, or any chronic and acute disease [23].
Hypoglycemia.
According to the American Diabetic Association (ADA), hypoglycemia proposes a threshold of < 70 mg/dl or 3.9 mmol/L [24].
Child weight was measured by an electronic digital weight scale for children who were comfortable being measured alone, but for children who were unable to be measured alone, we used mother-child weight combined measurement, and then the mother’s weight was measured to calculate the child’s weight through subtraction. The weight measurement was recorded to the nearest 0.1 kg. A child weight-for-age Z-score was classified as z-scores < -3 and >=-3 of the WHO standard curve [25].
Status of admission.
The consciousness level during admission as consciousness and unconsciousness [21].
Sampling procedure and sampling technique
We used the data source, the sample size of 590 from the previous study conducted at Debre Tabor Comprehensive Hospital [21]. Thus, for more information about the sampling procedure and sampling technique, please refer to this published article [21].
Data collection procedures and data quality assurance
Further analysis of the previously collected data was used. But during the primary data collection, the authors used interviews to administer questionnaires to collect the data from the participants) by trained nurses. The tool was adapted from west Bengal and India [23,26–29] and modified for the Ethiopian context. The treatment failure status of the children was obtained from the patient chart using the operational definition. To assure the validity of the tool, the English version of the questionnaire was translated to Amharic, which is the local language of the area, then back to English. A test was also conducted among 5% (30 participants) of the sample size outside of the study area to avoid information contamination. One-day intensive training was given for data collectors and supervisors to collect reliable and validated data. For the purpose of the current research, the data were accessed on July 20, 2023, from the primary authors of the previously published work [30] and the last author of the current study.
Data processing and analysis
The data were entered into Epi-data version 4.6.02 and exported to STATA software version 17 for cleaning, recoding, missing data checking, and further analysis. There was no missing data in the collected data, and the descriptive analysis was summarized by frequency and percentage. Using a p-value of less than 0.25 as a cutoff point, simple (bi-variable) binary logistic regression analysis was conducted to select the candidate variables for multivariable analysis. After the candidate variables were selected, a multivariable binary logistic regression analysis (backward step-wise logistic regression) was conducted. Variables with a p-value < 0.05 were declared significant factors. After repeated runs of the model, a model with the calibration (the assumption of Hosmer and Lemishow test), the better discriminatory power of area under the curve (AUC) in the receiver operating characteristics curve (ROC curve), a model with a small number of variables that best fit (parsimony model selection), plausibility, and easily interpretability were the guiding rules to select and develop a prediction model for treatment failure of SCAP. After the model was developed, its sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were analyzed. The AUC was also used to assess the discriminatory power and performance of the prediction model. A nomogram tool was developed based on prognostic determinants found significant in the prediction model. The bootstrapping technique (1000 random samples with replacement) was analyzed to evaluate the internal validity of the developed prediction model. The sensitivity, specificity, positive predictive value, negative predictive value, beta coefficients, and AUC were compared before and after bootstrapping, and then the optimism coefficient was evaluated. The developed nomogram model can estimate the risk of treatment failure for individual children. A decision curve analysis was used to assess the clinical and public health significance of the nomogram [28]. Risk classification was also analyzed (as a low-risk versus high-risk risk of treatment failure) based on the maximum Youden index with its corresponding optimal cutoff point.
Ethical declaration
Ethical approval and consent to participate.
Previously, during primary data collection, ethical approval was obtained from the Debre Tabor University ethical review board with the protocol number Dtu/RP/186/15, and written informed consent was obtained from each mother or legal guardian. But for the current research, we used the previously collected data. The owner of the data was Mr. Amare Kassaw, who was the primary author of the previously published work [21] and the last author of the current study.
Results
Baseline characteristics
About 580 participants, with a response rate of 98.3%, participated in the study. About 312 (46.2%) of them were male. Ninety-one (15.69%) and 74 (12.79%) participants were impaired (unconscious) and infected with HIV, respectively. Regarding the history of previous illness, 141 (24.31%) of under-five children have had a history of previous SCAP. Furthermore, 51 (8.79) of them were already infected with tuberculosis and known at the baseline (Table 1).
[Figure omitted. See PDF.]
Incidence of treatment failure and prognostic predictors of SCAP
The overall incidence of SCAP treatment failure among children who were admitted to Debre Tabor comprehensive specialized hospital was 28.1% (95% CI: 24.7%, 30.8%). After conducting the bi-variable analysis one by one, a total of 15 predictors were the candidate variables for further multivariable analysis (Table 2).
[Figure omitted. See PDF.]
Predictive model for SCAP treatment failure
Using the 17 predictors like pulmonary tuberculosis, meningitis, diarrhea, HIV status, dehydration, plural effusion, anaemia, hypoglycemia, measles, chest indrowing, weight for age, admission status, pertussis, pulse rate, sex of child, age of child, and previous SCAP, the multivariable logistic regression was analyzed, and finally, predictors like anaemia, previous history of SCAP, plural effusion, pulse rate, HIV status, measles, and chest indrowing were remained and used for the reduced model (Table 3). Finally, the nomogram was developed using the reduced model. The AUC of the original model was 0.7719 (95%CI: 0.729, 0.815) (Fig 1), and the calibration/model fitness at an angle of 450 showed that there was no difference between the predicted probability and the observed probability with a p-value of 0.917 (Fig 2). To estimate the performance of the original model, the optimal cut point was determined by the maximum Youden index value. Thus, the Youden index value was 0.4389 (Supporting File 1). At these cutoff points, the model had an accuracy of 76.38% (95% CI: 73%–80%), a sensitivity of 55.83% (95% CI: 48%–64%), a specificity of 84.41% (95% CI: 81%–88%), a positive predictive value (PPV) of 58.33% (95% CI: 59%–66%), and a negative predictive value of 83.02% (95% CI: 79%–86%). To evaluate the internal validity of the model, to avoid over interpreting, and to reduce the optimistic result from the original model, bootstrapping (1000 bootstrap samples with replacement) was used, and thus the AUC for the internally validated model was 0.7714 (0.728, 0.82) with an optimism coefficient of 0.0005 (Fig 1). The calibration of the internally validated model showed that there was agreement between the predicted and observed probability with a p-value of 0.549 (Fig 3). The very small optimism coefficient of a well-calibrated model can be applied to this model for the new sample. The risk prediction of SCAP treatment failure for the internally validated model using linear predictors.can be stated as follows:
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
P (SCAP treatment failure) = 1/e-(-2.67 + 1.14 * HIV status yes + 0.765 * plural effusion (yes) + 0.55 * anemia (yes) + 0.647 * measles (yes) + 0.915 * chest indrowing (yes) + 1.07 * history of SCAP (yes) + 0.603 * pulse rate (abnormal) (Table 3).
Nomogram for prediction of SCAP treatment failure
The nomogram was developed using a reduced model using variables with better predictive performance, biological plausibility, and easy interpretation. To develop the nomogram, seven predictors, such as anaemia, the previous history of SCAP, pulse rate, HIV status, measles, plural effusion, and chest indrowing, were included. The role of each predictor was also presented in the form of a nomogram division score. Using the total score of the SCAP treatment failure in the nomogram, the probability of treatment failure for each individual child can be calculated (Fig 4).
[Figure omitted. See PDF.]
Treatment failure risk classification using nomogram
To classify the risk of treatment failure, the maximum Youden index (max J = 0.4389) was used. At this value, the corresponding cut-off point was 0.3585. Using this cutoff point, the risk was classified as low risk for treatment failure (<0.3585) and high risk for treatment failure (≥0.3585). Thus, 126 (48.3%) of the patients were in the high-risk treatment failure group (Table 4).
[Figure omitted. See PDF.]
Decision curve analysis
The standardized net benefit analysis was evaluated using decision curve analysis, and thus, across the probability, the nomogram had the better net benefit (Fig 5).
[Figure omitted. See PDF.]
Discussion
In the current study, an attempt has been made to derivate and validate the model to predict the treatment failure associated with SCAP among under-five children who were admitted to Debre Tabor Comprehensive Specialized Hospital. This paper is studying the same domain in the same setting for the prediction of mid-term outcomes. The overarching objective would still be the same for selecting patients that require more intensive monitoring. To achieve this objective, a total of 580 under-5 children who were admitted to the Debre Tabor specialized comprehensive hospital during the follow-up period were considered for analysis. Thus, the incidence of SCAP treatment failure among under-five children was 28.1% (95% CI: 24.7%, 30.8%). The finding was higher than a study conducted in India (18.2%) [20]. This might be because of the difference in socio-economic status, health care access, and health-seeking behaviour between the two countries. The maximum Youden index value with its corresponding cut-off point was the guiding rule to classify the risk of treatment failure for SCAP, and thus 48.3% of patients were at high risk of SCAP treatment failure, which means at most 48.3% of the treatment failures may be identified earlier when using the prediction tool, followed by more intense monitoring of the high-risk group. At a value of the Youden index (max J = 0.4389) with its corresponding cut-off point of < 0.3585, it was classified as low risk for treatment failure (not performing closing monitoring), or about 11.6% of children with SCAP had a lower risk of treatment failure for not performing closing monitoring in pediatrics SCAP patients in the intensive care unit. The risk prediction model for SCAP treatment failure was developed using seven predictors: anaemia, previous history of SCAP, plural effusion, pulse rate, HIV status, measles, and chest indrowing. The discriminatory power of the model was evaluated using AUC, and its value was 0.7719 (95% CI: 0.729, 0.815) (Fig 1). The calibration or model fitness at an angle of 450 showed that there was no difference between the predicted probability and the observed probability with a p-value of 0.917 (Fig 2). This well-calibrated model can avoid the risk of overestimating or underestimating the outcome of interest. To validate the model, to avoid over interpreting, and to reduce the optimistic result from the original model, the bootstrapping technique (1000 bootstrap samples with replacement) was used, and the adjusted AUC for the validated model was 0.7714 (0.728, 0.82) with an optimism coefficient of 0.0005 (Fig 1). The calibration of the validated model also showed that there was agreement between the predicted and observed probability with a p-value of 0.549 (Fig 3), and the β coefficients of the bootstrapped model were nearly similar to those of the original model. Which implies that with the limited optimism coefficient (0.0005), our model can be well performed in a new sample. As far as our searching is concerned, there has been no research conducted on the derivation and validation of a risk prediction model for SCAP treatment failure among under-five children in Ethiopia, and thus it is hard to compare the findings with a model developed in the Ethiopian setting. Even though the model was slightly different in terms of the included predictors and biomarkers like white blood cells, lactate, procalcitonin, cortisol, B-type natriuretic peptide, and C-reactive protein, the performance of our model was in line with a study conducted in a tertiary care university hospital in Portugal between December 2009 and March 2012 (AUC = 0.81, 95% CI: 0.74-0.91) [31]. Using biomarkers for risk prediction of SCAP treatment failure is very important for clinical decisions about treatment failure associated with SCAP, but in less developed countries like Ethiopia, the applicability and measurability of these biomarkers are very difficult in routine health care. Therefore, in the current study, the model considers the most applicable predictors in the Ethiopian context.
Our model has lower sensitivity (79%), higher specificity (71%), an inline positive predictive value (60%), and a negative predictive value (86%) than the previously developed model at the tertiary care university hospital in Portugal [31]. Developing a treatment failure SCAP prediction tool and classifying patients based on their risk of treatment failure is crucial to clinical decision-making by reducing the treatment failure rate among under-five children, but there is no tool to achieve this objective. Hereafter, it is essential for the evaluation and early identification of patients for their risk of SCAP treatment failure. To do this, our model was developed using predictors like previous history (SCAP), abnormal pulse rate, HIV status, anaemia, plural effusion, measles, and chest indrowing. These predictors are easily measurable and practicable for clinical decision-making in low-infrastructure countries like Ethiopia. Furthermore, our model has a better net benefit across the threshold probabilities. Which means our model has clinical and public health importance after external validation. Using easily measurable predictors in the routine health care system that suits Ethiopia was the strength of the study, but our model needs external validation before being used for clinical decisions. Assessment of clinical failure by the treating physician could be prone to information bias. Thus, it would be important for a future validation study to assess treatment failure independently, preferably by someone blinded to clinical characteristics at baseline. Additionally, the lack of related articles makes it more difficult to compare in detail with the other findings.
Conclusion
The incidence of treatment failure among children with severe community-acquired pneumonia was high in Debre Tabor comprehensive hospital. The previous history of severe community-acquired pneumonia, abnormal pulse rate, chest indrowing, anemia, HIV status, and plural effusion were the significant factors to develop the predictive model. The model had good discriminatory performance and internally valid. Similarly, the model has a good calibration ability with an insignificant loss of accuracy from the original. The models can have the potential to improve treatment outcomes in the clinical settings. But needs external validation before use
Supporting information
Supportive File 1. Cut of point classifier.
https://doi.org/10.1371/journal.pone.0320448.s001
(XLSX)
STROBE Statement. Checklist of items that should be included in reports of observational studies.
https://doi.org/10.1371/journal.pone.0320448.s002
(DOCX)
Acknowledgments
The authors would like to give thanks to Debre Tabor University and Debre Tabor specialized comprehensive hospital.
References
1. 1. Harris M, et al. British thoracic society guidelines for the management of community acquired pneumonia in children: update 2011. Thorax. 2011;66(Suppl 2):ii1–23.
* View Article
* Google Scholar
2. 2. Menendez R, Torres A. Treatment failure in community-acquired pneumonia. Chest. 2007;132(4):1348–55. pmid:17934120
* View Article
* PubMed/NCBI
* Google Scholar
3. 3. Lutfiyya MN, Henley E, Chang LF, Reyburn SW. Diagnosis and treatment of community-acquired pneumonia. Am Fam Physician. 2006;73(3):442–50. pmid:16477891
* View Article
* PubMed/NCBI
* Google Scholar
4. 4. File TM Jr, Marrie TJ. Burden of community-acquired pneumonia in North American adults. Postgrad Med. 2010;122(2):130–41. pmid:20203464
* View Article
* PubMed/NCBI
* Google Scholar
5. 5. Torres A, Martin-Loeches I, Menendez RJI c. m. Research in community-acquired pneumonia: the next steps. 2017;43(9):1395–7.
6. 6. Álvarez-Lerma F, Torres AJCoicc. Severe community-acquired pneumonia. 2004;10(5):369-374.
7. 7. de Castro FR, Torres A. Optimizing treatment outcomes in severe community-acquired pneumonia. Am J Respir Med. 2003;2(1):39–54. pmid:14720021
* View Article
* PubMed/NCBI
* Google Scholar
8. 8. Ortqvist A, Kalin M, Lejdeborn L, Lundberg B. Diagnostic fiberoptic bronchoscopy and protected brush culture in patients with community-acquired pneumonia. Chest. 1990;97(3):576–82. pmid:2306961
* View Article
* PubMed/NCBI
* Google Scholar
9. 9. Arancibia F, Ewig S, Martinez JA, Ruiz M, Bauer T, Marcos MA, et al. Antimicrobial treatment failures in patients with community-acquired pneumonia: causes and prognostic implications. Am J Respir Crit Care Med. 2000;162(1):154–60. pmid:10903235
* View Article
* PubMed/NCBI
* Google Scholar
10. 10. Yealy DM, Auble TE, Stone RA, Lave JR, Meehan TP, Graff LG, et al. The emergency department community-acquired pneumonia trial: Methodology of a quality improvement intervention. Ann Emerg Med. 2004;43(6):770–82. pmid:15159711
* View Article
* PubMed/NCBI
* Google Scholar
11. 11. Low DE, Mazzulli T, Marrie TJC. Progressive and nonresolving pneumonia. J Clinical Microbiology. 2005;11(3):247–52.
* View Article
* Google Scholar
12. 12. Viasus D, et al. Early, short and long-term mortality in community-acquired pneumonia. 2018;2(5):10.
* View Article
* Google Scholar
13. 13. Begg EJ, Barclay ML, Kirkpatrick CMJB. The therapeutic monitoring of antimicrobial agents. J Clinical Pharmacy. 2001;52(S1):35–43.
* View Article
* Google Scholar
14. 14. Milkovich G. Intravenous-to-oral transition therapy in community-acquired pneumonia: the INOVA Health System experience. Pharmacotherapy. 2001;21(7 Pt 2):83S-88S. pmid:11446523
* View Article
* PubMed/NCBI
* Google Scholar
15. 15. Weingarten SR, Riedinger MS, Varis G, Noah MS, Belman MJ, Meyer RD, et al. Identification of low-risk hospitalized patients with pneumonia. Implications for early conversion to oral antimicrobial therapy. Chest. 1994;105(4):1109–15. pmid:8162734
* View Article
* PubMed/NCBI
* Google Scholar
16. 16. Siegel RE, Halpern NA, Almenoff PL, Lee A, Cashin R, Greene JG. A prospective randomized study of inpatient iv. antibiotics for community-acquired pneumonia. The optimal duration of therapy. Chest. 1996;110(4):965–71. pmid:8874253
* View Article
* PubMed/NCBI
* Google Scholar
17. 17. Ramirez JA, Vargas S, Ritter GW, Brier ME, Wright A, Smith S, et al. Early switch from intravenous to oral antibiotics and early hospital discharge: a prospective observational study of 200 consecutive patients with community-acquired pneumonia. Arch Intern Med. 1999;159(20):2449–54. pmid:10665893
* View Article
* PubMed/NCBI
* Google Scholar
18. 18. Genné D, Sommer R, Kaiser L, Saaïdia A, Pasche A, Unger PF, et al. Analysis of factors that contribute to treatment failure in patients with community-acquired pneumonia. Eur J Clin Microbiol Infect Dis. 2006;25(3):159–66. pmid:16528540
* View Article
* PubMed/NCBI
* Google Scholar
19. 19. Ewig S, Torres A, Angeles Marcos M, Angrill J, Rañó A, de Roux A, et al. Factors associated with unknown aetiology in patients with community-acquired pneumonia. Eur Respir J. 2002;20(5):1254–62. pmid:12449182
* View Article
* PubMed/NCBI
* Google Scholar
20. 20. Sabapathy S. Predictors of treatment failure in severe pneumonia treated with Ampicillin in tertiary care center. 2019.
21. 21. Kassaw A, Kerebih G, Zeleke S, Chanie ES, Dessalegn N, Bante B, et al. Survival status and predictors of mortality from severe community-acquired pneumonia among under-five children admitted at Debre Tabor comprehensive specialized hospital: a prospective cohort study. Front Pediatr. 2023;11:1141366. pmid:37346893
* View Article
* PubMed/NCBI
* Google Scholar
22. 22. Khawaja A, et al. Etiology and outcome of severe community acquired pneumonia in immunocompetent adults. 2013;13(X):1–6.
23. 23. Mengist B, Tesfa M, Kassie B. Time to recovery and predictors of severe community-acquired pneumonia among pediatric patients in Debre Markos referral hospital, North West Ethiopia: A retrospective follow-up study. PLoS One. 2020;15(9):e0239655. pmid:32976491
* View Article
* PubMed/NCBI
* Google Scholar
24. 24. Care DJDC. Glycemic targets: standards of medical care in diabetes—2019. Diabetes Care. 2019;42(Supplement 1):S61-70.
* View Article
* Google Scholar
25. 25. Tibebu NS, et al. Underweight and its associated factors among children 6–59 months of age in Debre Tabor Town, Amhara Region of Ethiopia, 2019: A community-based cross-sectional study. Journal Name. 2020:469–76.
* View Article
* Google Scholar
26. 26. Keleb A, Sisay T, Alemu K, Ademas A, Lingerew M, Kloos H, et al. Pneumonia remains a leading public health problem among under-five children in peri-urban areas of north-eastern Ethiopia. PLoS One. 2020;15(9):e0235818. pmid:32915807
* View Article
* PubMed/NCBI
* Google Scholar
27. 27. Chakraborty S, Mukherjee A, Bhattacherjee S, Majumdar R, Chatterjee M, Dasgupta S. Risk factors for pneumonia mortality in under-five children in a tertiary care hospital of Darjeeling district of West Bengal: A prospective case-control study. Indian J Public Health. 2020;64(4):368–73. pmid:33318387
* View Article
* PubMed/NCBI
* Google Scholar
28. 28. Negash AA, et al. Bacteremic community-acquired pneumonia in Ethiopian children: etiology, antibiotic resistance, risk factors, and clinical outcome. Open Forum Infectious Diseases. Oxford University Press US; 2019.
29. 29. Ramachandran P, Nedunchelian K, Vengatesan A, Suresh S. Risk factors for mortality in community acquired pneumonia among children aged 1-59 months admitted in a referral hospital. Indian Pediatr. 2012;49(11):889–95. pmid:22791667
* View Article
* PubMed/NCBI
* Google Scholar
30. 30. Kirolos A, Blacow RM, Parajuli A, Welton NJ, Khanna A, Allen SJ, et al. The impact of childhood malnutrition on mortality from pneumonia: a systematic review and network meta-analysis. BMJ Glob Health. 2021;6(11):e007411. pmid:34848440
* View Article
* PubMed/NCBI
* Google Scholar
31. 31. Pereira JM, et al. Early prediction of treatment failure in severe community-acquired pneumonia: The PRoFeSs score. 2019;53:38–45.
Citation: Agimas MC, Tesfie TK, Derseh NM, Kassaw A (2025) Derivation and validation of a model to predict treatment failure among under five children with severe community acquired pneumonia who are admitted at Debre Tabor specialized comprehensive hospital. PLoS ONE 20(3): e0320448. https://doi.org/10.1371/journal.pone.0320448
About the Authors:
Muluken Chanie Agimas
Roles: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing
E-mail: [email protected]
Affiliation: Department of Epidemiology and Biostatistics, Institute of Public Health, College of Medicine and Health Science, University of Gondar, Gondar, Ethiopia
ORICD: https://orcid.org/0000-0003-0808-124X
Tigabu Kidie Tesfie
Roles: Conceptualization, Formal analysis
Affiliation: Department of Epidemiology and Biostatistics, Institute of Public Health, College of Medicine and Health Science, University of Gondar, Gondar, Ethiopia
ORICD: https://orcid.org/0000-0002-8250-125X
Nebiyu Mekonnen Derseh
Roles: Data curation, Software, Visualization, Writing – original draft
Affiliation: Department of Epidemiology and Biostatistics, Institute of Public Health, College of Medicine and Health Science, University of Gondar, Gondar, Ethiopia
ORICD: https://orcid.org/0000-0003-1355-2763
Amare Kassaw
Roles: Conceptualization, Software, Supervision
Affiliation: Department of Pediatrics and Child Health Nursing, College of Health sciences, Debre Tabor University, Debre Tabor, Ethiopia
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
1. Harris M, et al. British thoracic society guidelines for the management of community acquired pneumonia in children: update 2011. Thorax. 2011;66(Suppl 2):ii1–23.
2. Menendez R, Torres A. Treatment failure in community-acquired pneumonia. Chest. 2007;132(4):1348–55. pmid:17934120
3. Lutfiyya MN, Henley E, Chang LF, Reyburn SW. Diagnosis and treatment of community-acquired pneumonia. Am Fam Physician. 2006;73(3):442–50. pmid:16477891
4. File TM Jr, Marrie TJ. Burden of community-acquired pneumonia in North American adults. Postgrad Med. 2010;122(2):130–41. pmid:20203464
5. Torres A, Martin-Loeches I, Menendez RJI c. m. Research in community-acquired pneumonia: the next steps. 2017;43(9):1395–7.
6. Álvarez-Lerma F, Torres AJCoicc. Severe community-acquired pneumonia. 2004;10(5):369-374.
7. de Castro FR, Torres A. Optimizing treatment outcomes in severe community-acquired pneumonia. Am J Respir Med. 2003;2(1):39–54. pmid:14720021
8. Ortqvist A, Kalin M, Lejdeborn L, Lundberg B. Diagnostic fiberoptic bronchoscopy and protected brush culture in patients with community-acquired pneumonia. Chest. 1990;97(3):576–82. pmid:2306961
9. Arancibia F, Ewig S, Martinez JA, Ruiz M, Bauer T, Marcos MA, et al. Antimicrobial treatment failures in patients with community-acquired pneumonia: causes and prognostic implications. Am J Respir Crit Care Med. 2000;162(1):154–60. pmid:10903235
10. Yealy DM, Auble TE, Stone RA, Lave JR, Meehan TP, Graff LG, et al. The emergency department community-acquired pneumonia trial: Methodology of a quality improvement intervention. Ann Emerg Med. 2004;43(6):770–82. pmid:15159711
11. Low DE, Mazzulli T, Marrie TJC. Progressive and nonresolving pneumonia. J Clinical Microbiology. 2005;11(3):247–52.
12. Viasus D, et al. Early, short and long-term mortality in community-acquired pneumonia. 2018;2(5):10.
13. Begg EJ, Barclay ML, Kirkpatrick CMJB. The therapeutic monitoring of antimicrobial agents. J Clinical Pharmacy. 2001;52(S1):35–43.
14. Milkovich G. Intravenous-to-oral transition therapy in community-acquired pneumonia: the INOVA Health System experience. Pharmacotherapy. 2001;21(7 Pt 2):83S-88S. pmid:11446523
15. Weingarten SR, Riedinger MS, Varis G, Noah MS, Belman MJ, Meyer RD, et al. Identification of low-risk hospitalized patients with pneumonia. Implications for early conversion to oral antimicrobial therapy. Chest. 1994;105(4):1109–15. pmid:8162734
16. Siegel RE, Halpern NA, Almenoff PL, Lee A, Cashin R, Greene JG. A prospective randomized study of inpatient iv. antibiotics for community-acquired pneumonia. The optimal duration of therapy. Chest. 1996;110(4):965–71. pmid:8874253
17. Ramirez JA, Vargas S, Ritter GW, Brier ME, Wright A, Smith S, et al. Early switch from intravenous to oral antibiotics and early hospital discharge: a prospective observational study of 200 consecutive patients with community-acquired pneumonia. Arch Intern Med. 1999;159(20):2449–54. pmid:10665893
18. Genné D, Sommer R, Kaiser L, Saaïdia A, Pasche A, Unger PF, et al. Analysis of factors that contribute to treatment failure in patients with community-acquired pneumonia. Eur J Clin Microbiol Infect Dis. 2006;25(3):159–66. pmid:16528540
19. Ewig S, Torres A, Angeles Marcos M, Angrill J, Rañó A, de Roux A, et al. Factors associated with unknown aetiology in patients with community-acquired pneumonia. Eur Respir J. 2002;20(5):1254–62. pmid:12449182
20. Sabapathy S. Predictors of treatment failure in severe pneumonia treated with Ampicillin in tertiary care center. 2019.
21. Kassaw A, Kerebih G, Zeleke S, Chanie ES, Dessalegn N, Bante B, et al. Survival status and predictors of mortality from severe community-acquired pneumonia among under-five children admitted at Debre Tabor comprehensive specialized hospital: a prospective cohort study. Front Pediatr. 2023;11:1141366. pmid:37346893
22. Khawaja A, et al. Etiology and outcome of severe community acquired pneumonia in immunocompetent adults. 2013;13(X):1–6.
23. Mengist B, Tesfa M, Kassie B. Time to recovery and predictors of severe community-acquired pneumonia among pediatric patients in Debre Markos referral hospital, North West Ethiopia: A retrospective follow-up study. PLoS One. 2020;15(9):e0239655. pmid:32976491
24. Care DJDC. Glycemic targets: standards of medical care in diabetes—2019. Diabetes Care. 2019;42(Supplement 1):S61-70.
25. Tibebu NS, et al. Underweight and its associated factors among children 6–59 months of age in Debre Tabor Town, Amhara Region of Ethiopia, 2019: A community-based cross-sectional study. Journal Name. 2020:469–76.
26. Keleb A, Sisay T, Alemu K, Ademas A, Lingerew M, Kloos H, et al. Pneumonia remains a leading public health problem among under-five children in peri-urban areas of north-eastern Ethiopia. PLoS One. 2020;15(9):e0235818. pmid:32915807
27. Chakraborty S, Mukherjee A, Bhattacherjee S, Majumdar R, Chatterjee M, Dasgupta S. Risk factors for pneumonia mortality in under-five children in a tertiary care hospital of Darjeeling district of West Bengal: A prospective case-control study. Indian J Public Health. 2020;64(4):368–73. pmid:33318387
28. Negash AA, et al. Bacteremic community-acquired pneumonia in Ethiopian children: etiology, antibiotic resistance, risk factors, and clinical outcome. Open Forum Infectious Diseases. Oxford University Press US; 2019.
29. Ramachandran P, Nedunchelian K, Vengatesan A, Suresh S. Risk factors for mortality in community acquired pneumonia among children aged 1-59 months admitted in a referral hospital. Indian Pediatr. 2012;49(11):889–95. pmid:22791667
30. Kirolos A, Blacow RM, Parajuli A, Welton NJ, Khanna A, Allen SJ, et al. The impact of childhood malnutrition on mortality from pneumonia: a systematic review and network meta-analysis. BMJ Glob Health. 2021;6(11):e007411. pmid:34848440
31. Pereira JM, et al. Early prediction of treatment failure in severe community-acquired pneumonia: The PRoFeSs score. 2019;53:38–45.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2025 Agimas et al. This is an open access article distributed under the terms of the Creative Commons Attribution License: http://creativecommons.org/licenses/by/4.0/ (the “License”), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
Background
Severe community-acquired pneumonia related treatment failure is persistence of features of severe pneumonia after initiation of antimicrobial therapy or a worsening clinical condition within 48–72 hours of the commencement of the antibiotics. Even though it is the most devastating public health problem in Ethiopia, there is no study to derivate and validate a model to predict treatment failure. To do this, nomogram was used to estimate the probability of treatment failure for each individual child and to classify their risk of treatment failure.
Objective
to develop and validate the model to predict treatment failure among under five children with severe community-acquired pneumonia in Debre Tabor comprehensive specialized hospital.
Method
A secondary analysis of the previously collected prospective follow-up study was used for further analysis among 590 under-5 children hospitalized with severe community-acquired pneumonia. The STATA version 17 software was used for analysis. Descriptive analysis was summarized by frequency and percentage. A multivariable binary logistic regression was also conducted, and the model performance was evaluated using the receiver operating characteristics curve with its area under the curve and calibration curve. Internal validation of the model was assessed using the bootstrap technique. The decision curve analysis was also used to evaluate the usefulness of the nomogram.
Results
The incidence of treatment failure among severe community-acquired pneumonia children was 28.1% (95% CI: 24.7%–30.8%). The previous history of severe community-acquired pneumonia, abnormal pulse rate, chest indrowing, anemia, HIV status, and plural effusion remained for the final model. The area under the curve for the original model and validated model was 0.7719 (95%CI: 0.729, 0.815) and 0.7714 (95% CI: 0.728–0.82), respectively. The decision curve analysis showed that the nomogram had a better net benefit across the threshold probability.
Conclusion
The incidence of treatment failure among children with severe community-acquired pneumonia was high in Debre Tabor comprehensive hospital. The previous history of severe community-acquired pneumonia, abnormal pulse rate, chest indrowing, anemia, HIV status, and plural effusion were the significant factors to develop the predictive model. The model had good discriminatory performance and internally valid. Similarly, the model has a good calibration ability with an insignificant loss of accuracy from the original. The models can have the potential to improve treatment outcomes in the clinical settings. But needs external validation before use.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details
; Tesfie, Tigabu Kidie
; Nebiyu Mekonnen Derseh
; Kassaw, Amare