Comparative Analysis of Parametric Survival Models in HIV Patient Data
DOI:
https://doi.org/10.6000/Keywords:
Parametric survival analysis, Gompertz distribution, Weibull model, Lomax distribution, Exponential survival model, HIV/AIDS, Antiretroviral therapy, Maximum likelihood estimationAbstract
This study explores the efficacy of four key parametric survival models-Weibull, Gompertz, Lomax, and Exponential-in assessing mortality risk among HIV-positive patients undergoing antiretroviral therapy (ART). The research examined a retrospective cohort of 2,794 individuals, noting 124 deaths (4.4%) and 2,670 censored cases (95.6%), utilizing time-to-event data. Each model was estimated using maximum likelihood estimation (MLE) and assessed using various model selection criteria, including the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC). The Gompertz distribution emerged as the best fit (AIC = 45,943.33; BIC = 45,961.58), followed by the Weibull model, while the Lomax and Exponential models showed higher AIC/BIC values and less stable fits. The optimized parameters for the Gompertz model were determined as λ = 0.00316 and α = 1.77x10-6, indicating a gradually increasing hazard rate over time. Model adequacy was further confirmed using Cox-Snell residuals (via Nelson-Aalen cumulative hazard) and Cox-Snell residual Q-Q plots for diagnostic evaluation. The Gompertz model demonstrated the highest coefficient of determination (R2 = 0.9817), followed by the Weibull (R2 = 0.9168), while the Lomax and Exponential models both had lower R2 values (0.5989), underscoring the superior predictive capability of the Gompertz model. Additionally, Cox proportional hazards regression identified significant mortality predictors, such as age at ART initiation (HR = 1.05, p < 0.001), male sex (HR = 1.60, p < 0.01), and last recorded body weight (HR = 0.94, p < 0.001). In contrast, baseline CD4 count and WHO stage were not significant. The model’s concordance index (C = 0.85) indicated high predictive accuracy. This study is motivated by the ongoing variability in HIV survival outcomes despite the extensive use of ART. By comparing these parametric models, the research enhances the understanding of mortality dynamics, aiding clinicians and policymakers in selecting optimal model structures for precise survival prediction, improved ART program monitoring, and informed patient management.These findings highlight significant clinical implications for HIV care, identifying age at ART initiation, male sex, and lower body weight as mortality predictors,indicating where targeted actions are needed. The Gompertz model’s superior performance offers a robust method for the prediction of long-term survival, underlining the need for monitoring comorbidities and the management of treatment-related side effects. With this model, HIV programs will be better positioned to flag high-risk patients, time interventions more appropriately, and allocate resources to reduce preventable deaths among their aging populations.
References
[1]Md Faiyazuddin, et al. The impact of artificial intelligence on healthcare: a comprehensive review of advancements in diagnostics, treatment, and operational efficiency. In: Health Science Reports 2025; 8(1): e70312. https://doi.org/10.1002/hsr2.70312[2]Trickey A, Zhang L, Sabin CA, Sterne JAC. Life expectancy of people with HIV on long-term antiretroviral therapy in Europe and North America: a cohort study. In: The Lancet Healthy Longevity 2022; 3: S2. https://doi.org/10.1016/S2666-7568(22)00063-0[3]Althoff KN, et al. The shifting age distribution of people with HIV using antiretroviral therapy in the United States. In: Aids 2022; 36(3): 459-471. https://doi.org/10.1097/QAD.0000000000003128[4]Teklehaymanot AN, Lemma TB, Gudina EK, Getnet M, Amdisa D, Dadi LS. Predictors of Mortality among Adult People Living with HIV and Its Implications for Appointment Spacing Model Approach Care in Southwest Ethiopia. In: 2020. https://doi.org/10.21203/rs.3.rs-35093/v1[5]Kumssa TH, Mulu A, Mihret A, Asfaw ZG. Competing risks multi-state model for time-to-event data analysis of HIV/AIDS: a retrospective cohort national datasets, Ethiopia. In: BMC Infectious Diseases 2024; 24(1): 1412. https://doi.org/10.1186/s12879-024-10280-9[6]Korenromp EL, Williams BG, Schmid GP, Dye C. Clinical prognostic value of RNA viral load and CD4 cell counts during untreated HIV-1 infection—a quantitative review. In: PloS One 2009; 4(6): e5950. https://doi.org/10.1371/journal.pone.0005950[7]Dessie ZG, Zewotir T, Mwambi H, North D. Modelling of viral load dynamics and CD4 cell count progression in an antiretroviral naive cohort: using a joint linear mixed and multistate Markov model. In: BMC infectious diseases 2020; 20(1): 246. https://doi.org/10.1186/s12879-020-04972-1[8]Haushona N, Esterhuizen TM, Thabane L, Machekano R. An empirical comparison of time-to-event models to analyse a composite outcome in the presence of death as a competing risk. In: Contemporary Clinical Trials Communications 2020; 19: 100639. https://doi.org/10.1016/j.conctc.2020.100639[9]Payne CF, Houle B, Chinogurei C, Herl CR, Kabudula CW, Kobayashi LC, Salomon JA, Manne-Goehler J. Differences in healthy longevity by HIV status and viral load among older South African adults: an observational cohort modelling study. In: The Lancet HIV 2022; 9(10): e709-e716. https://doi.org/10.1016/S2352-3018(22)00198-9[10]Glaubius R, et al. Disease progression and mortality with untreated HIV infection: Evidence synthesis of HIV seroconverter cohorts, antiretroviral treatment clinical cohorts and population-based survey data. In: Journal of the International AIDS Society 2021; 24: e25784. https://doi.org/10.1002/jia2.25784[11]Mustefa YA, Chen D-G. Accelerated failure-time model with weighted least-squares estimation: application on survival of HIV positives. In: Archives of Public Health 2021; 79(1): 88. https://doi.org/10.1186/s13690-021-00617-0[12]Omer ME, Mustafa M, Ali N, Rahman NHA. Non-Mixture Cure Model Estimation in Bladder Cancer Patients: A Novel Approach with Exponentiated Weibull Exponential Distribution. In: Asian Pacific Journal of Cancer Prevention: APJCP 2023; 24(12): 4167. https://doi.org/10.31557/APJCP.2023.24.12.4167[13]Hussain S, Rashid MS, Ul Hassan M, Ahmed R. The generalized exponential extended exponentiated family of distributions: Theory, properties, and applications. In: Mathematics 2022; 10(19): 3419. https://doi.org/10.3390/math10193419[14]Skalski JR, Whitlock SL. Vitality models found useful in modeling tag-failure times in acoustic-tag survival studies. In: Animal Biotelemetry 2020; 8(1): 26. https://doi.org/10.1186/s40317-020-00213-z[15]Dzinza R, Ngwira A. Comparing parametric and Cox regression models using HIV/AIDS survival data from a retrospective study in Ntcheu district in Malawi. In: Journal of Public Health Research 2022; 11(3): 22799036221125328. https://doi.org/10.1177/22799036221125328[16]Delva W, Eaton JW, Meng F, Fraser C, White RG, Vickerman P, Boily M-C, Hallett TB. HIV treatment as prevention: optimising the impact of expanded HIV treatment programmes. In: PLoS Medicine 2012; 9(7): e1001258. https://doi.org/10.1371/journal.pmed.1001258[17]Onyekwere CK, Nwankwo CK, Abonongo J, Asogwa EC, Shafiq A. Economic growth dynamics: a machine learning-augmented nonlinear autoregressive distributed lag model of
Comparative Analysis of Parametric Survival Models in HIV Patient DataInternational Journal of Statistics in Medical Research, 2025, Vol. 14945asymmetric effect. In: Innovation in Computer and Data Sciences 2025; 1(1): 9-31. https://doi.org/10.64389/icds.2025.01125[18]Onyekwere CK, Nwankwo CK, Obulezi OJ, Ezeilo CI. The featurevalue paradox: Unsupervised discovery of strategic archetypes in the smartphone market using machine learning. In: Journal of Artificial Intelligence in Engineering Practice 2025; 2(2): 65-72. https://doi.org/10.21608/jaiep.2025.420689.1024[19]Onyekwere CK, Nwankwo CK, Apameh DG. A Hybrid Machine Learning Framework for Multi-Objective Performance Optimization and Anomaly Detection in Maritime Operations. In: Innovation in Computer and Data Sciences 2026; 2(1). https://doi.org/10.64389/icds.2026.02131[20]Asogwa EC, Nwankwo MP, Oguadimma EE, Okechukwu CP, Suleiman AA. Hybrid LSTM-CNN deep learning framework for stock price prediction with google stock and reddit sentiment data. In: Innovation in Computer and Data Sciences 2025; 1(1): 32-50. https://doi.org/10.64389/icds.2025.01126[21]Ugbor G, Jamal F, Khan S, Shawki AW. Generative AI for drug discovery: Accelerating molecular design with deep learning using Nigerian local content. In: Innovation in Computer and Data Sciences 2025; 1(1): 66-77. https://doi.org/10.64389/icds.2025.01128[22]Nnaekwe K, Ani E, Obieke V, Okechukwu C, Usman A, Othman M. Forecasting seasonal rainfall with time series, machine learning and deep learning. In: Innovation in Computer and Data Sciences 2025; 1(1): 51-65. https://doi.org/10.64389/icds.2025.01127[23]Asogwa EC, Okechukwu OM, Charles EI, Tochukwu B. A Medical Intelligent Process Model Using Ontology Based Technique. In: International Journal of Simulation: Systems, Science & Technology 2024; 25(1). https://doi.org/10.5013/IJSSST.a.25.01.03[24]Madut DB, Park LP, Yao J, Reddy EA, Njau B, Ostermann J, Whetten K, Thielman NM. Predictors of mortality in treatment experienced HIV-infected patients in northern Tanzania. In: Plos One 2020; 15(10): e0240293. https://doi.org/10.1371/journal.pone.0240293[25]de la Mora L, Mallolas J, Ambrosioni J. Epidemiology, treatment and prognosis of HIV infection in 2024: A practical review. In: Medicina ClÅLınica (English Edition) 2024; 162(11): 535-541. https://doi.org/10.1016/j.medcle.2023.12.010[26]Little RJA, Rubin DB. Statistical analysis with missing data. John Wiley & Sons 2019. https://doi.org/10.1002/9781119482260[27]Pedersen AB, Mikkelsen EM, Cronin-Fenton D, Kristensen NR, Pham TM, Pedersen L, Petersen I. Missing data and multiple imputation in clinical epidemiological research. In: Clinical Epidemiology 2017; 157-166. https://doi.org/10.2147/CLEP.S129785[28]Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM, Carpenter JR. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. In: BMJ 2009; 338. https://doi.org/10.1136/bmj.b2393[29]Cohen JE, Bohk-Ewald C, Rau R. Gompertz, Makeham, and Siler models explain Taylor’s law in human mortality data. In: Demographic Research 2018; 38: 773-841. https://doi.org/10.4054/DemRes.2018.38.29[30]Ramaki Z, Alizadeh M, Tahmasebi S, Afshari M, Contreras-Reyes JE, Yousof HM. TheWeighted FlexibleWeibull Model: Properties, Applications, and Analysis for Extreme Events. In: Mathematical and Computational Applications 2025; 30(2): 42. https://doi.org/10.3390/mca30020042[31]Jafari AA, Tahmasebi S. Gompertz-power series distributions. In: Communications in Statistics-Theory and Methods 2016; 45(13): 3761-3781. https://doi.org/10.1080/03610926.2014.911904[32]Witten M, Satzer W. Gompertz survival model parameters: Estimation and sensitivity. In: Applied Mathematics Letters 1992; 5(1): 7-12. https://doi.org/10.1016/0893-9659(92)90125-S[33]Wilson DL. A comparison of methods for estimating mortality parameters from survival data. In: Mechanisms of Ageing and Development 1993; 66(3): 269-281. https://doi.org/10.1016/0047-6374(93)90014-I[34]Albrecher HN, Bladt M, Bladt M. Multivariate fractional phase-type distributions. In: Fractional Calculus and Applied Analysis 2020; 23(5): 1431-1451. https://doi.org/10.1515/fca-2020-0071[35]Hussam E, Almetwally EM. Statistical inference on the perk distribution under progressive-stress Type-II Adaptive Progressive Hybrid Censoring. In: 2024. https://doi.org/10.21203/rs.3.rs-5361936/v1[36]Di Palo C. On a closed-form expression and its approximation to Gompertz life disparity. In: Demographic Research 2023; 49: 1-12. https://doi.org/10.4054/DemRes.2023.49.1[37]Li H, Tan KS, Tuljapurkar S, Zhu W. Gompertz law revisited: Forecasting mortality with a multi-factor exponential model. In: Insurance: Mathematics and Economics 2021; 99: 268-281. https://doi.org/10.1016/j.insmatheco.2021.03.018[38]Lin GD, Dou X, Kuriki S. The bivariate lack-of-memory distributions. In: Sankhya A 2019; 81(2): 273-297. https://doi.org/10.1007/s13171-017-0119-1[39]Rajitha CS, Akhilnath A. Generalization of the Lindley distribution with application to COVID-19 data: Rajitha CS and A. Akhilnath. In: International Journal of Data Science and Analytics 2025; 20(2): 291-311. https://doi.org/10.1007/s41060-022-00369-2[40]Osman AA, Esse AA, Muse AH. Analyzing factors affecting age at first birth among married women in Somalia: a Bayesian shared frailty modelling approach using SDHS 2020. In: BMC Women’s Health 2025; 25(1): 346. https://doi.org/10.1186/s12905-025-03900-2[41]Kearns B, Stevens J, Ren S, Brennan A. How uncertain is the survival extrapolation? A study of the impact of different parametric survival models on extrapolated uncertainty about hazard functions, lifetime mean survival and cost effectiveness. In: Pharmacoeconomics 2020; 38(2): 193-204. https://doi.org/10.1007/s40273-019-00853-x[42]Alnssyan B, Ahmad Z, Malela-Majika J-C, Seong J-T, Shafik W. On the identifiability and statistical features of a new distributional approach with reliability applications. In: AIP Advances 2023; 13(12). https://doi.org/10.1063/5.0178555[43]Tolba AH, Muse AH, Fayomi A, Baaqeel HM, Almetwally EM. The Gull Alpha Power Lomax distributions: Properties, simulation, and applications to modeling COVID-19 mortality rates. In: Plos One 2023; 18(9): e0283308. https://doi.org/10.1371/journal.pone.0283308[44]Al-Essa LA, Abdel-Hamid AH, Alballa T, Hashem AF. Reliability analysis of the triple modular redundancy system under step-partially accelerated life tests using Lomax distribution. In: Scientific Reports 2023; 13(1): 14719. https://doi.org/10.1038/s41598-023-41363-3[45]Ferreira PH, Ramos E, Ramos PL, Gonzales JFB, Tomazella VLD, Ehlers RS, Silva EB, Louzada F. Objective Bayesian analysis for the Lomax distribution. In: Statistics & Probability Letters 2020; 159: 108677. https://doi.org/10.1016/j.spl.2019.108677
946International Journal of Statistics in Medical Research, 2025, Vol. 14Elkalzah et al.[46]Benkhelifa L. Modi linear failure rate distribution with application to survival time data. In: arXiv preprint 2025; arXiv:2509.20831. [47]Imran M, Alsadat N, Tahir MH, Jamal F, Elgarhy M, Ahmad H, Johannssen A. The development of an extended Weibull model with applications to medicine, industry and actuarial sciences. In: Scientific Reports 2024; 14(1): 12338. https://doi.org/10.1038/s41598-024-61308-8[48]Nkomo W, Oluyede B, Chipepa F. Type I heavy-tailed family of generalized Burr III distributions: properties, actuarial measures, regression and applications. In: Statistics in Transition New Series 2025; 26(1): 93-115. https://doi.org/10.59139/stattrans-2025-006[49]Mansournia MA, Geroldinger A, Greenland S, Heinze G. Separation in logistic regression: causes, consequences, and control. In: American Journal of Epidemiology 2018; 187(4): 864-870. https://doi.org/10.1093/aje/kwx299[50]Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. In: Statistics in Medicine 2006; 25(1): 127-141. https://doi.org/10.1002/sim.2331[51]Sabin CA. Do people with HIV infection have a normal life expectancy in the era of combination antiretroviral therapy? In: BMC Medicine 2013; 11(1): 251. https://doi.org/10.1186/1741-7015-11-251[52]Crowther MJ, Look MP, Riley RD. Multilevel mixed effects parametric survival models using adaptive Gauss-Hermite quadrature with application to recurrent events and individual participant data meta-analysis. In: Statistics in Medicine 2014; 33(22): 3844-3858. https://doi.org/10.1002/sim.6191[53]Shoko C, Chikobvu D. A superiority of viral load over CD4 cell count when predicting mortality in HIV patients on therapy. In: BMC Infectious Diseases 2019; 19(1): 169. https://doi.org/10.1186/s12879-019-3781-1[54]Engsig FN, et al. Long-term mortality in HIV-positive individuals virally suppressed for. 3 years with incomplete CD4 recovery. In: Clinical Infectious Diseases 2014; 58(9): 1312-1321. [55]Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. In: American Journal of Epidemiology 2007; 165(6): 710-718. https://doi.org/10.1093/aje/kwk052[56]Greenland S, Mansournia MA. Penalization, bias reduction, and default priors in logistic and related categorical and survival regressions. In: Statistics in Medicine 2015; 34(23): 3133-3143. https://doi.org/10.1002/sim.6537[57]Brennan AT, Maskew M, Sanne I, Fox MP. The interplay between CD 4 cell count, viral load suppression and duration of antiretroviral therapy on mortality in a resource-limited setting. In: Tropical Medicine & International Health 2013; 18(5): 619-631. https://doi.org/10.1111/tmi.12079[58]Harvey NC, McCloskey E, Kanis JA, Compston J, Cooper C. Cost-effective but clinically inappropriate: new NICE intervention thresholds in osteoporosis (Technology Appraisal 464). In: Osteoporosis International 2018; 29(7): 1511-1513. https://doi.org/10.1007/s00198-018-4505-x[59]Tekelehaimanot AN, Belachew T, Gudina EK, Getnet M, Amdisa D, Dadi LS. Predictors of mortality among adult people living with HIV and its implications for appointment spacing model approach care. In: Ethiopian Journal of Health Sciences 2021; 31(5). https://doi.org/10.4314/ejhs.v31i5.3
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Policy for Journals/Articles with Open Access
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are permitted and encouraged to post links to their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work
Policy for Journals / Manuscript with Paid Access
Authors who publish with this journal agree to the following terms:
- Publisher retain copyright .
- Authors are permitted and encouraged to post links to their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work .
