The DOI matches its Crossref record; 58 of the 58 references it cites were found in a registry or an index.
Interpretation of tests of heterogeneity and bias in meta-analysis
Abstract
Statistical tests of heterogeneity and bias, in particular publication bias, are very popular in meta-analyses. These tests use statistical approaches whose limitations are often not recognized. Moreover, it is often implied with inappropriate confidence that these tests can provide reliable answers to questions that in essence are not of statistical nature. Statistical heterogeneity is only a correlate of clinical and pragmatic heterogeneity and the correlation may sometimes be weak. Similarly, statistical signals may hint to bias, but seen in isolation they cannot fully prove or disprove bias in general, let alone specific causes of bias, such as publication bias in particular. Both false-positive and false-negative signals of heterogeneity and bias can be common and their prevalence may be anticipated based on some rational considerations. Here I discuss the major common challenges and flaws that emerge in using and interpreting statistical tests of heterogeneity and bias in meta-analyses. I discuss misinterpretations that can occur at the level of statistical inference, clinical/pragmatic inference and specific cause attribution. Suggestions are made on how to avoid these flaws, use these tests properly and learn from them.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“These limitations are common in meta-analyses of cross-sectional studies (Ioannidis, 2008).”
From Decision making capacity for treatment in psychiatric inpatients: a systematic review and meta-analysis (2024) · cites this paper as Ioannidis, 2008
Citations
Cites
Measuring inconsistency in meta-analyses
Higgins, J. P. T., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003) · BMJ · doi:10.1136/bmj.327.7414.557 · Cited by 63,139
Bias in meta-analysis detected by a simple, graphical test
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997) · BMJ · doi:10.1136/bmj.315.7109.629 · Cited by 56,722
Quantifying heterogeneity in a meta‐analysis
Higgins, J. P. T., & Thompson, S. G. (2002) · Statistics in Medicine · doi:10.1002/sim.1186 · Cited by 36,763
Operating characteristics of a rank correlation test for publication bias
Begg, C. B., & Mazumdar, M. (1994) · Biometrics · doi:10.2307/2533446 · Cited by 16,955
Evidence based medicine: what it is and what it isn't
Sackett, D. L., Rosenberg, W. M. C., Gray, J. A. M., Haynes, R. B., & Richardson, W. S. (1996) · BMJ · doi:10.1136/bmj.312.7023.71 · Cited by 15,769
Why Most Published Research Findings Are False
Ioannidis, J. P. A. (2005) · PLoS Medicine · doi:10.1371/journal.pmed.0020124 · Cited by 10,505
Self-citationThe combination of estimates from different experiments
Cochran, W. G. (1954) · Biometrics · doi:10.2307/3001666 · Cited by 5,380
Why Most Published Research Findings Are False
Ioannidis, J. P. A. (2005) · CHANCE · doi:10.1080/09332480.2005.10722754 · Cited by 4,702
Self-citationAssessing heterogeneity in meta-analysis: Q statistic or I² index?
Huedo-Medina, T. B., Sánchez-Meca, J., Marín-Martínez, F., & Botella, J. (2006) · Psychological Methods · doi:10.1037/1082-989x.11.2.193 · Cited by 4,256
Funnel plots for detecting bias in meta-analysis
Sterne, J. A. C., & Egger, M. (2001) · Journal of Clinical Epidemiology · doi:10.1016/s0895-4356(01)00377-8 · Cited by 3,779
Cited by
Sorted by citations
Items 401–500 of 695
The relationship between self-reflection and mental health: a meta-analysis review
He, W., & Gan, J. (2025) · Current Psychology · doi:10.1007/s12144-025-07415-9 · Cited by 6
“Finally, publication bias was assessed using Egger’s repression test (Egger et al., 1997 ), Begg’s rank correlation test (Ioannidis, 2008 ), and Duval and Tweedie’s trim and fill analysis (Aitkin, 1999 ), using the metabias and trimfill functions from the meta package.”
Cai, P., Ramadan, M., Mei, C., & Hadfield, K. (2025) · Advances in Mental Health · doi:10.1080/18387357.2025.2522435 · Cited by 7
Demeke, A., Bitacha, G. K., Gebrie, H., Tefera, S., Getu, E., Dube, S. J., Bayable, A., Kassie, A., Muleta, B., Eba, W. W., et al. (2025) · BMC Infectious Diseases · doi:10.1186/s12879-025-11568-0 · Cited by 6
The Effect of Liver Cirrhosis on Patients Undergoing Cardiac Surgery
Liu, F., Li, Z., Liu, X., Liu, X.-Y., & Yang, J. (2023) · Global Heart · doi:10.5334/gh.1270 · Cited by 6
Kitaw, T. A., Getie, ፟., Asgedom, S. G., Adisu, M. A., Tilahun, B. D., Zemariam, A. B., Alamaw, A. W., Faris, A. M., Habtie, T. E., Munie, M. A., et al. (2024) · Global Epidemiology · doi:10.1016/j.gloepi.2024.100180 · Cited by 6
How to teach the fundamentals of meta-analyses
Brusselaers, N. (2015) · Annals of Epidemiology · doi:10.1016/j.annepidem.2015.08.004 · Cited by 6
“The Cochran Q test and I2 test both evaluate statistical heterogeneity [43].”
MethodCephalometric norms of skeletal relationship among populations in selected Arab countries
Al‐Shammery, D., Almubarak, S., Hezaim, A., Alkhunein, R., Pani, S. C., & Mossadomi, H. (2016) · Saudi Journal of Oral Sciences · doi:10.4103/1658-6816.188079 · Cited by 6
Zhang, X., Jiang, Z.-Y., Zhang, Z., Chen, H., Wu, K., He, J., & Xie, C. (2020) · Drug Design Development and Therapy · doi:10.2147/dddt.s229621 · Cited by 6
Xie, Y., Fang, X., Hua, H., & Zhou, P. (2022) · Frontiers in Pharmacology · doi:10.3389/fphar.2022.812085 · Cited by 6
“The I2 test was performed to evaluate the heterogeneity of the studies and ranged from 0 to 100%; I2 values of 25 and 50% were used as cutoffs for modest and high heterogeneity (Ioannidis, 2008).”
MethodWatanabe, T., Sieg, M., Lunde, S. J., Persson, M. L., Taneja, P., Baad‐Hansen, L., Pigg, M., & Vase, L. (2023) · Journal of Oral Rehabilitation · doi:10.1111/joor.13414 · Cited by 6
Liu, X., Tong, Y. A., Li, Z., Liu, F., Liu, X., Zhang, W., & Peng, D. (2023) · International Journal of Colorectal Disease · doi:10.1007/s00384-023-04372-2 · Cited by 6
“I2 value and the chi-squared test were used to assess the statistical heterogeneity [37, 38].”
Gavgani, L. F., Laghousi, D., Sarbakhsh, P., Jahangiri, L., Vahed, N., & Hajebrahimi, S. (2023) · Journal of Research in Medical Sciences · doi:10.4103/jrms.jrms_1114_21 · Cited by 6
“Due to the different designs between studies, the random effect model was used for meta‐analysis.[9] Meta‐regression test and sensitivity analysis were used to identify the reasons for heterogeneity.”
Lei, L., Wang, J., Zhang, L., Chen, Y., Yuan, P., & Liu, D. (2017) · Oncotarget · doi:10.18632/oncotarget.20803 · Cited by 6
“On the other hand, we observed large degree of heterogeneities among included studies and we also try to well interpret the causes from different aspects [30].”
BackgroundAntoniou, E. E., Gelbke, H., Ballach, J., & Zeegers, M. P. (2020) · Toxicology Research and Application · doi:10.1177/2397847319899080 · Cited by 6
“The I(2) (Isquared statistic) was used to describe the percentage of variation across studies due to heterogeneity.(25) The association between DMF exposure and liver toxicity in exposed compared to non-exposed workers was investigated firstly, in a dose–response subgroup analysis, and secondly, in an overall analysis…”
MethodHolper, L. (2020) · Frontiers in Psychiatry · doi:10.3389/fpsyt.2020.00633 · Cited by 7
Statistische Methoden zur Detektion und Adjustierung von Publikationsbias
Schwarzer, G., & Rücker, G. (2010) · Zeitschrift für Evidenz Fortbildung und Qualität im Gesundheitswesen · doi:10.1016/j.zefq.2010.03.016 · Cited by 6
LGE and the Risk of Sudden Death in HCM
Alla, V. M., Koneru, S., Hunter, C. B., & Mooss, A. N. (2012) · JACC. Cardiovascular imaging · doi:10.1016/j.jcmg.2012.05.004 · Cited by 6
Prevalence of Schistosoma mansoni infection in Ethiopia: A systematic review and meta-analysisPreprint
Hussen, S., Assegu, D., & Shimelis, T. (2019) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/610113 · Cited by 6
“05) [18].”
How Identity Fusion Predicts Extreme Pro-Group Orientations: A Meta-AnalysisPreprint
Varmann, A. H., Kruse, L., Bierwiaczonek, K., Gómez, Á., Vázquez, A., & Kunst, J. R. (2022) · doi:10.31234/osf.io/prasc · Cited by 6
Investigating the Heterogeneity of “Study Twins”
Röver, C., & Friede, T. (2024) · Biometrical Journal · doi:10.1002/bimj.202300387 · Cited by 15
“Significance tests for homogeneity generally are known to have little power, especially for as few as 𝑘 = 2 studies (Hardy and Thompson 1998; Ioannidis 2008; Pereira et al. 2010), and heterogeneity often tends to be underestimated (Kontopantelis, Springate, and Reeves 2013).”
Getachew, A., Molla, M., Admasie, A., & Azage, M. (2024) · BMC Pediatrics · doi:10.1186/s12887-024-05084-4 · Cited by 8
“The heterogeneity of studies was quantified using the I-squared statistic [18].”
MethodSafety and efficacy of the enhanced recovery after surgery protocol in hepatectomy for liver cancer
Lv, Q., Xiang, Y.-C., Qiu, Y.-Y., & Xiang, Z. (2024) · Clinics and Research in Hepatology and Gastroenterology · doi:10.1016/j.clinre.2024.102493 · Cited by 5
Factors informing tax compliance: A meta-analytical study
Reddy, D., Moodley, A., & Olugbara, O. O. (2024) · South African Journal of Economic and Management Sciences · doi:10.4102/sajems.v27i1.5738 · Cited by 5
Li, C., Liu, C., Peng, C. H., Li, T., Xie, B., & Liu, Z. (2025) · Ecological Informatics · doi:10.1016/j.ecoinf.2025.103352 · Cited by 5
Abate, B. B., Tusa, B. S., Sendekie, A. K., Temesgen, D., Mekuria, K., Alamaw, A. W., Azmeraw, M., Zemariam, A. B., Kitaw, T. A., Kassaw, A., et al. (2025) · PLoS ONE · doi:10.1371/journal.pone.0334852 · Cited by 6
Dawood, Z. S., Brown, Z. J., Endo, Y., Katayama, E., Munir, M. M., Alaimo, L., Ruff, S. M., Lima, H. A., Woldesenbet, S., & Pawlik, T. M. (2023) · Surgical Oncology · doi:10.1016/j.suronc.2023.102013 · Cited by 5
Tiongco, R. E., Cayanan, N. D., Catacata, M., & Dominguez, M. J. (2024) · Biomarkers · doi:10.1080/1354750x.2024.2326538 · Cited by 5
“Publication bias testing using the egger regression asymmetry test and the Begg–Mazumdar correlation test was performed on associations with ≥10 studies and had significant results (ioannidis 2008; Dalton et al .”
MethodKitaw, T. A., Abate, B. B., Tilahun, B. D., Yilak, G., & Haile, R. N. (2024) · Health Science Reports · doi:10.1002/hsr2.2159 · Cited by 5
Methods for assessing inverse publication bias of adverse events
Xing, X., Xu, C. Q., Amer, F. M. A., Shi, L., Zhu, J., & Lin, L. (2024) · Contemporary Clinical Trials · doi:10.1016/j.cct.2024.107646 · Cited by 5
Shu, X., Lv, Q., Li, Z., Liu, F., Liu, X., Li, L., Tong, Y. A., Liu, X., Wang, C., Peng, D., et al. (2024) · European journal of medical research · doi:10.1186/s40001-024-01995-1 · Cited by 5
“The I 2 value and the Chi-squared test were used to evaluate the heterogeneity of identified research [24, 25].”
MethodStatistical Heterogeneity in Oral Health Meta-Analyses
Tatas, Z., Kyriakou, E., Seehra, J., Pandis, N., & Spineli, L. M. (2025) · Journal of Dental Research · doi:10.1177/00220345251316279 · Cited by 7
“Ideally, the estimated I 2 should always be reported with a 95% CI (Ioannidis et al. 2007; Ioannidis 2008; von Hippel 2015).”
Makrygiannakis, M. A., Athanasiou, C. A., & Kaklamanos, E. G. (2022) · European Journal of Orthodontics · doi:10.1093/ejo/cjac052 · Cited by 5
Yang, J. B., Liu, Q., Bai, Y., Zhao, H., He, T., Zhao, Z., Huang, M., Jiang, M., Zhang, R., & Zhang, M. (2023) · Frontiers in Oncology · doi:10.3389/fonc.2022.1025855 · Cited by 5
“0) 33/26 27/24 2714 46(14-103) * pN0 and IH”
MethodHabib, A., Athanasiou, A. E., Makrygiannakis, M. A., & Kaklamanos, E. G. (2023) · European Journal of Orthodontics · doi:10.1093/ejo/cjac076 · Cited by 6
Aydıngöz, S. E., Teimoori, A., Orhan, H. G., Efe, O. E., Kibaroğlu, S., & ERDEM, Ş. R. (2023) · European Journal of Pharmacology · doi:10.1016/j.ejphar.2023.175564 · Cited by 5
Rajkumar, B., Parameswaran, R., Sanjana, M., Boovaraghavan, S., & Vijayalakshmi, D. (2023) · Indian Journal of Dental Research · doi:10.4103/ijdr.ijdr_338_22 · Cited by 5
Gebremichael, B., Lassi, Z. S., Begum, M., & Zhou, S. J. (2024) · Clinical Nutrition ESPEN · doi:10.1016/j.clnesp.2024.06.029 · Cited by 5
Agimas, M. C., Derseh, N. M., Aragaw, F. M., Abebe, M. T., Birhan, T. Y., & Alemayehu, M. A. (2024) · PLoS ONE · doi:10.1371/journal.pone.0303940 · Cited by 5
Li, Y., Quan, X., Zhou, C., Duan, X., Nie, M., & Si, H. (2025) · Journal of Orthopaedic Surgery and Research · doi:10.1186/s13018-025-05694-3 · Cited by 5
“Guidelines for the I [2] test state: 0 ∼ 25% is low-level heterogeneity, 25 ∼ 50% is moderate heterogeneity, 50 ∼ 75% is high-level heterogeneity, and more than 75% is considerable level heterogeneity [16].”
Posti, J. P., Salo, P. P., Ruohonen, S., Valve, L., Muszkat, M., Sofowora, G. G., Kurnik, D., Stein, C. M., Perola, M., Scheinin, M., et al. (2013) · Pharmacogenetics and Genomics · doi:10.1097/fpc.0b013e32835d247f · Cited by 5
Quality of meta-analyses for randomized trials in the field of hypertension
Roush, G. C., Amante, B., Singh, T., Ayele, H., Araoye, M., Yang, D., Kostis, W. J., Elliott, W. J., Kostis, J. B., & Berlin, J. A. (2016) · Journal of Hypertension · doi:10.1097/hjh.0000000000001094 · Cited by 5
“It can be argued that a test for publication bias is not valid if there is substantial heterogeneity [39]; among meta-analyses with an I(2) of less than 50%, 59% failed to conduct a test for publication bias.”
BackgroundSule, I. O., Mendelsohn, A. B., & Panas, R. M. (2022) · Health Services Research and Managerial Epidemiology · doi:10.1177/23333928221131709 · Cited by 5
“However, it is not reasonable to deduce that all the heterogeneity should be elucidated because the residual heterogeneity is expected to be recognized in the statistical analysis.(33) In such a manner, assessing these covariates in each study is impractical.”
BackgroundKunze, K. N., Kay, J., Pareek, A., Dahmen, J., Nwachukwu, B. U., Williams, R. J., KARLSSON, J. L., & de SA, D. L. (2023) · Knee Surgery Sports Traumatology Arthroscopy · doi:10.1007/s00167-023-07328-9 · Cited by 6
“Understanding these particular aspects of the meta-analytic process is essential to ensure (1) that data are being appropriately handled and (2) determine whether the results presented may be generalisable and reliable [12, 25].”
BackgroundPeng, D., Cheng, Y.-X., & Liao, G. (2021) · World Journal of Surgical Oncology · doi:10.1186/s12957-021-02230-5 · Cited by 5
“The value of I(2) and the results of the chi-squared test were used to assess the statistical heterogeneity [20, 21].”
MethodAntidepressants for the treatment of depression in patients with cancer
Ostuzzi, G., Matcham, F., Dauchy, S., Barbui, C., & Hotopf, M. (2014) · Cochrane Database of Systematic Reviews · doi:10.1002/14651858.cd011006 · Cited by 4
Genome-wide association and epidemiological evidence linking air pollution to aging mechanisms
Liu, X., Peng, D., Liu, X., Sheng, Y.-H., Zhao, J.-N., Xu, R.-B., Zhang, D., & Chen, B. (2025) · Journal of Hazardous Materials · doi:10.1016/j.jhazmat.2025.139273 · Cited by 5
Tiongco, R. E., Arbas, R., Caguiat, A., Castro, E. J., Cruz, K. C. D., Dominguez, M. J., Guintu, M. A. A. D., Mercado, J. A., & Vallo, J. (2025) · Personalized Medicine · doi:10.1080/17410541.2025.2544352 · Cited by 5
Akkapelli, S. S., Boorla, V. S., & Majooju, M. (2025) · Nutrition Reviews · doi:10.1093/nutrit/nuaf283 · Cited by 4
Vera-Ponce, V. J., Ballena-Caicedo, J., Valladolid-Sandoval, L. A. M., & Montoya, F. E. Z. (2026) · International Journal of Statistics in Medical Research · doi:10.6000/1929-6029.2026.15.01 · Cited by 4
Li, J., Xu, Y., Tan, S.-D., & Wang, Z. (2024) · Medicine · doi:10.1097/md.0000000000038475 · Cited by 4
“[23,24] Groups with an I² >50% indicated high heterogeneity, prompting the use of a random effects model; a P value < .1 was considered statistically significant for these analyses.”
MethodWang, J., Bai, Z., & Chen, B. (2024) · Annals of Vascular Surgery · doi:10.1016/j.avsg.2024.04.025 · Cited by 4
Zhang, B., Shen, Y., Jin, L., & Wang, J. (2024) · Langenbeck s Archives of Surgery · doi:10.1007/s00423-024-03406-9 · Cited by 4
“Interstudy heterogeneity was assessed using the Cochran Q statistic (chi-square test) [ 16 ], along with the I 2 statistic as a measure describing degrees of heterogeneity [ 17 ].”
Gebremichael, B., Lassi, Z. S., Begum, M., & Zhou, S. J. (2024) · Maternal and Child Nutrition · doi:10.1111/mcn.13737 · Cited by 4
“The I 2 values of ≤ 25, 50 – 74, and ≥ 75% indicated low, moderate, and high heterogeneity (Ioannidis, 2008).”
Ortiz, H. A. D. L. B., Parizotto, N. A., Lange, C. C., & Liebano, R. E. (2025) · Journal of Hand Therapy · doi:10.1016/j.jht.2024.10.001 · Cited by 4
Badru, O. A., Edeh, J. C., Okonkwo, R. I., Bain, L. E., & Adeagbo, O. (2025) · AIDS Research and Treatment · doi:10.1155/arat/5757907 · Cited by 4
“Low, moderate, and high heterogeneity was defned as I 2 of ≤ 25%, ≤ 50%, and ≥ 75%, respectively [39].”
Is the recurrence rate higher in obese patients undergoing inguinal hernia surgery?
Li, Z., Song, M., Liu, J., Jiang, B., Hu, W., & Zheng, X. (2025) · Hernia · doi:10.1007/s10029-025-03301-2 · Cited by 4
“If I 2 > 50%, it was considered high het - erogeneity, the random effect model was used and p 0.1 was considered statistically significant [24 ].”
InfluentialPrevalence of human brucellosis in Ethiopia: systematic review and meta-analysis
Asrie, F., Birhan, M., Dagnew, M., & Berhane, N. (2025) · BMC Infectious Diseases · doi:10.1186/s12879-025-10502-8 · Cited by 4
“Significant heterogeneity was considered for P < 0.10 and I 2 > 50% [40, 41].”
InfluentialRosa, G. R. M. L., Giovane, C. D., Pedullá, E., Kowalski, J., Chapple, I., Minozzi, S., Amaliya, A., & Polosa, R. (2025) · Systematic Reviews · doi:10.1186/s13643-025-02839-7 · Cited by 4
“To determine the potential influence of effect modifiers on the outcomes, we will conduct a subgroup analysis [45].”
Tusa, B. S., Alati, R., Ayano, G., Betts, K. S., Weldesenbet, A. B., & Dachew, B. A. (2025) · BJPsych Open · doi:10.1192/bjo.2025.67 · Cited by 5
Prediction Models of Primary Membranous Nephropathy: A Systematic Review and Meta-Analysis
Geng, C., Huang, L., Li, Y., Wang, Y., Li, G., & Feng, Y. (2023) · Journal of Clinical Medicine · doi:10.3390/jcm12020559 · Cited by 4
“Statistical heterogeneity was estimated using I2 statistic [15], assessed as low if I2 of <25%, moderate if I2 ranged from 26% to 75%, and high if I2 of >75%.”
MethodRelative likelihood ratios for neutral comparisons of statistical tests in simulation studies
Huang, Q., & Trinquart, L. (2023) · Biometrical Journal · doi:10.1002/bimj.202200102 · Cited by 4
“…result does not exclude the possibility of small-study effects, in particular, publication bias; in fact, our simulation studies show that the negative likelihood ratios are moderate, indicating that a negative test result does not greatly decrease the prior probability of bias (Ioannidis, 2008).”
ResultWen, Z., Zhou, X., & Peng, D. (2023) · Medicine · doi:10.1097/md.0000000000035554 · Cited by 4
“The chi-squared test and the I2 value were conducted to calculated the heterogeneity of results.[16,29] Articles were considered with high heterogeneity when I2 > 50%, and I2 < 50% presented a low heterogeneity.”
MethodLiu, X., Zhang, B., Peng, D., Liu, F., Li, Z., & Wang, C. (2024) · Updates in Surgery · doi:10.1007/s13304-024-01864-w · Cited by 4
“The value of I 2 and the Chi-squared test were used to evaluate the statistical heterogeneity [29, 30].”
Traoré, K. A., Akapovi, M. M., Ouédraogo, N., Ouoba, J. B., Roques, P., & Barro, N. (2024) · BMC Infectious Diseases · doi:10.1186/s12879-024-09668-4 · Cited by 4
“The Cochrane Q test was used to evaluate the existence of heterogeneity, and p 0.05 was considered to indicate statistical significance [22].”
MethodInfluentialTiongco, R. E., Cayanan, N. D., Dominguez, M. J., Geonzon, K., Mercado, J. A., & Pineda-Cortel, M. R. (2025) · The Egyptian Journal of Internal Medicine · doi:10.1186/s43162-025-00402-3 · Cited by 4
Abuhay, H. W., Yenit, M. K., Melese, M., Alemu, G. G., & Aragaw, F. M. (2025) · PLoS ONE · doi:10.1371/journal.pone.0315529 · Cited by 4
“%) degrees of heterogeneity [24].”
Dudy, N. K., Verma, S., & Chitra, P. (2025) · Orthodontics and Craniofacial Research · doi:10.1111/ocr.12924 · Cited by 4
Jyothish, S., Athanasiou, A. E., Makrygiannakis, M. A., & Kaklamanos, E. G. (2021) · PLoS ONE · doi:10.1371/journal.pone.0247011 · Cited by 4
“10 used for the heterogeneity tests [35].”
AlWadiyah, M., Athanasiou, A. E., Makrygiannakis, M. A., & Kaklamanos, E. G. (2022) · International Orthodontics · doi:10.1016/j.ortho.2022.100638 · Cited by 4
Sim, G. G., See, A. H. M., & Quah, L. J. J. (2023) · The Journal of Trauma: Injury, Infection, and Critical Care · doi:10.1097/ta.0000000000004004 · Cited by 4
“21 Because of the inability to gather complete data from all study authors, it was not possible to conduct a meta-analysis for the secondary outcomes.”
BackgroundPrognostic Value of Postoperative Complication for Gastric Cancer
Ren, L.-F., Xu, Y., & Long, J.-G. (2024) · Journal of Laparoendoscopic & Advanced Surgical Techniques · doi:10.1089/lap.2023.0456 · Cited by 4
Is recovery enhancement after gastric cancer surgery really a safe approach for elderly patients?
Li, Z., Luo, X.-J., Liu, F., Liu, X., Shu, X., Tong, Y. A., Lv, Q., Liu, X., Zhang, W., & Peng, D. (2024) · World Journal of Gastrointestinal Oncology · doi:10.4251/wjgo.v16.i4.1334 · Cited by 4
“If I 2 > 50%, it was considered high heterogeneity, the random effect model was used and P 0.1 was considered statistically significant[21].”
InfluentialKitaw, T. A., Abate, B. B., Yilak, G., Tilahun, B. D., Faris, A. M., Walle, G. T., & Haile, R. N. (2024) · AIDS Research and Therapy · doi:10.1186/s12981-024-00630-7 · Cited by 4
“Zero invers variance (I 2 ) revealed true homogeneity, whereas 25%, 50% and 75% had low, moderate and high heterogeneity, respectively [14, 15].”
BackgroundGuide pratique de recherche en réadaptation
Tétreault, S., & Blais‐Michaud, S. (2014) · De Boeck Supérieur eBooks · doi:10.3917/dbu.guill.2014.01.0187 · Cited by 4
Huang, Q., Shu, H., Zeng, C., Qiu, P., Xiong, X., & Lu, X. (2022) · BMJ Open · doi:10.1136/bmjopen-2022-066903 · Cited by 4
Meta-Analysis of Therapeutic Trials
Cappelleri, J. C., Ioannidis, J. P. A., & Lau, J. (2018) · doi:10.1201/9781351110273-130001125 · Cited by 4
Self-citationJafari-Koulaee, A., Goudarzian, A. H., Beik, S., Heidari, T., & Hesamzadeh, A. (2020) · Asia Pacific Journal of Health Management · doi:10.24083/apjhm.v15i4.351 · Cited by 3
Herrera-Bennett, A., Heene, M., Lakens, D., & Ufer, S. (2020) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/zt3g9 · Cited by 5
Ren, S., Huang, L., Li, Y., & Feng, Y. (2024) · PLoS ONE · doi:10.1371/journal.pone.0311603 · Cited by 3
“Statistical heterogeneity was estimated using I 2 statistic [10].”
MethodHan, L., Wang, Y., Zhong, J., Zhang, M., Li, W., Zhou, Y., Yao, N., Zhang, L., Dou, X., & Wang, X. (2025) · Pediatric Surgery International · doi:10.1007/s00383-025-06004-y · Cited by 3
Abdunabi, F., Kowash, M., Khamis, A. H., & Abuzayeda, M. (2025) · Journal of International Society of Preventive and Community Dentistry · doi:10.4103/jispcd.jispcd_5_25 · Cited by 3
Zorzi, F., Scilipoti, P., Moretto, S., Gonzalez, C. G., Nannola, N., Robesti, D., Folcia, A., Chicaud, M., Kutchukian, S., Candela, L., et al. (2026) · Cancers · doi:10.3390/cancers18050821 · Cited by 3
“…heterogeneity across available studies, including differences in patient selection, tumor risk profile, endoscopic technique, outcome definitions, and follow-up duration, we deliberately adopted a descriptive, sample-size-weighted approach and refrained from formal meta-analytic pooling [13,39].”
Brini, S., & Leung, T. I. (2026) · Journal of Medical Internet Research · doi:10.2196/92132 · Cited by 3
Buhamer, S. N., Kaklamanos, E. G., Kowash, M., Hussein, I., Salami, A., & Halabi, M. A. (2021) · PLoS ONE · doi:10.1371/journal.pone.0259293 · Cited by 4
Liu, F., Wen, Z., Liu, X., Li, Z., Lv, Q., Zhang, W., & Peng, D. (2023) · World Journal of Surgical Oncology · doi:10.1186/s12957-023-03167-7 · Cited by 3
“We used the random effects model, and P < 0.05 was considered statistically significant [23].”
MethodInfluentialMengistie, B. A., Abiy, S. A., Anteneh, T. A., Wubneh, S. B., Andualem, F., Melese, M., Fentie, E. A., Setegn, A., Demeke, M., Amare, G. N., et al. (2025) · Systematic Reviews · doi:10.1186/s13643-025-02837-9 · Cited by 3
“Heterogeneity across the included studies was classified as low, moderate, or high when the values of I 2 were 25%, 50–75%, and 75%, respectively [19].”
Techane, M. A., Atalell, K. A., & Wondim, A. (2022) · International Journal of Africa Nursing Sciences · doi:10.1016/j.ijans.2022.100428 · Cited by 3
Al-Saqi, R. K., Athanasiou, A. E., Makrygiannakis, M. A., & Kaklamanos, E. G. (2023) · PLoS ONE · doi:10.1371/journal.pone.0285309 · Cited by 4
Liu, X., Wen, Z., Liu, F., Li, Z., Liu, X., Zhang, W., & Peng, D. (2024) · World Journal of Gastrointestinal Oncology · doi:10.4251/wjgo.v16.i1.133 · Cited by 3
“The chi-squared test and the I 2 value were used to evaluate the statistical heterogeneity[21,22].”
Zweygberg, A. L., Martin, F. Z., Brynedal, B., Lindholm, E. S., Kosidou, K., Ahlqvist, V. H., & Magnusson, C. (2024) · American Journal of Obstetrics and Gynecology · doi:10.1016/j.ajog.2024.02.015 · Cited by 3
Mañalac, A. R., Catacata, M., Mercado, J. A., Flake, C. C., Navarro, A. T., & Tiongco, R. E. (2024) · Human Gene · doi:10.1016/j.humgen.2024.201286 · Cited by 3
Tola, A., Assefa, N., Dessie, Y., Oljira, L., Regassa, L. D., Gure, T., & Gobena, T. (2024) · Systematic Reviews · doi:10.1186/s13643-024-02526-z · Cited by 4
“The I 2 values of 25%, 50%, and 75% represent low, medium, and substantial heterogeneity, respectively [32, 33].”
BackgroundA Problem of Too Much Heterogeneity
Downey, L. A., & Guzzetta, N. A. (2020) · Anesthesia & Analgesia · doi:10.1213/ane.0000000000004719 · Cited by 3
Zapf, A. (2018) · Springer eBooks · doi:10.1007/978-3-319-78966-8_11 · Cited by 3
Quantitative phenotype analysis to identify, validate and compare rat disease models
Zhao, Y., Smith, J. R., Wang, S., Dwinell, M. R., & Shimoyama, M. E. (2019) · Database · doi:10.1093/database/baz037 · Cited by 3
“For example, true heterogeneity in study population (due to subgroups with a different intervention effect) will lead to funnel plot asymmetry (36).”
BackgroundLi, J., Jiang, J., Jing, C., Zheng, W., & Hua-shan, P. (2020) · Medicine · doi:10.1097/md.0000000000022917 · Cited by 3
A ‘reluctant’ critical review: ‘Manual for evidence‐based clinical practice (2015)’
Seshia, S. S. (2015) · Journal of Evaluation in Clinical Practice · doi:10.1111/jep.12509 · Cited by 3
“Ioannidis suggests that when I(2) statistic = 0, ‘a common inference is that no heterogeneity exists, but this can be potentially misleading, especially if the uncertainty is large’ [61].”
BackgroundMethodTusa, B. S., Alati, R., Ayano, G., Betts, K. S., Weldesenbet, A. B., & Dachew, B. A. (2025) · Psychiatry Research · doi:10.1016/j.psychres.2025.116428 · Cited by 2
“Statistical heterogeneity was evaluated using the Q-test and I² statistic (Cowan, 2001; Ioannidis, 2008).”
Zhong, X.-M., & Liu, X. (2025) · Oncology Letters · doi:10.3892/ol.2025.15030 · Cited by 2
Prognostic importance of red blood cell distribution width in patients with glioma: A meta‑analysis
Chen, Y., Chen, F., Luo, Y.-X., & Xiong, L. (2025) · Oncology Letters · doi:10.3892/ol.2025.15217 · Cited by 2
“To evaluate the statistical heterogeneity, the I 2 value and the χ 2 test were used (24,25).”
Metadata
- Authors
- Ioannidis, J. P. A.
- Affiliations
- University of Ioannina (GR)
- Corresponding author
- Ioannidis, John P. A.
- Venue
- Journal of Evaluation in Clinical Practice
- Publisher
- Wiley
- Published
- 2008-10
- Volume
- 14
- Issue
- 5
- Pages
- 951-957
- ISSN
- 1356-1294 (ISSN-L)
- 1365-2753
- OpenAlex
- W2144112749
- Cited by
- 696
- References
- 57
- Citations by year
- 2026: 52 · 2025: 69 · 2024: 98 · 2023: 72 · 2022: 50 · 2021: 64 · 2020: 49 · 2019: 34 · 2018: 30 · 2017: 23 · 2016: 23 · 2015: 26 · 2014: 27 · 2013: 22 · 2012: 15
- FWCI
- 4.52
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Journal of Evaluation in Clinical Practice (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Keywords
- heterogeneity testing, publication bias, meta-analysis, statistical heterogeneity, clinical heterogeneity, false positive signals, bias detection
- MeSH terms
- Data Interpretation, Statistical, Humans, Regression Analysis, Research Design, Risk, Sensitivity and Specificity, Meta-Analysis as Topic, Reproducibility of Results, Bias, Confidence Intervals, Publication Bias, Evidence-Based Medicine
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
