Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2010 Aug 9:2:57-66.
doi: 10.2147/clep.s9242.

Interpreting meta-analysis according to the adequacy of sample size. An example using isoniazid chemoprophylaxis for tuberculosis in purified protein derivative negative HIV-infected individuals

Kristian Thorlund  1 , Aranka Anema, Edward Mills
Affiliations

Interpreting meta-analysis according to the adequacy of sample size. An example using isoniazid chemoprophylaxis for tuberculosis in purified protein derivative negative HIV-infected individuals

Kristian Thorlund et al. Clin Epidemiol. .

Abstract

Objective: To illustrate the utility of statistical monitoring boundaries in meta-analysis, and provide a framework in which meta-analysis can be interpreted according to the adequacy of sample size. To propose a simple method for determining how many patients need to be randomized in a future trial before a meta-analysis can be deemed conclusive.

Study design and setting: Prospective meta-analysis of randomized clinical trials (RCTs) that evaluated the effectiveness of isoniazid chemoprophylaxis versus placebo for preventing the incidence of tuberculosis disease among human immunodeficiency virus (HIV)-positive individuals testing purified protein derivative negative. Assessment of meta-analysis precision using trial sequential analysis (TSA) with LanDeMets monitoring boundaries. Sample size determination for a future trials to make the meta-analysis conclusive according to the thresholds set by the monitoring boundaries.

Results: The meta-analysis included nine trials comprising 2,911 trial participants and yielded a relative risk of 0.74 (95% CI, 0.53-1.04, P = 0.082, I(2) = 0%). To deem the meta-analysis conclusive according to the thresholds set by the monitoring boundaries, a future RCT would need to randomize 3,800 participants.

Conclusion: Statistical monitoring boundaries provide a framework for interpreting meta-analysis according to the adequacy of sample size and project the required sample size for a future RCT to make a meta-analysis conclusive.

Keywords: adequacy of sample size; human immunodeficiency virus (HIV); isoniazid chemoprophylaxis; meta-analysis; purified protein derivative negative; randomized clinical trials (RCTs); trial sequential analysis (TSA); tuberculosis.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Forest plot of the individual trial RR estimates and 95% CIs, and the DerSimonian-Laird pooled RR estimate. Notes: RR estimates smaller than 1.00 favor use of isoniazid chemoprophylaxis (IHZ) for prevention of tuberculosis in purified protein derivative negative HIV-infected individuals, RR estimates larger than 1.00 favor controls.
Figure 2
Figure 2
Trial sequential analysis (TSA) monitoring boundaries for moderate and strong evidence (the downward sloping lines) in the current meta-analysis. Notes: The monitoring boundaries were constructed using the Lan-DeMets alpha-spending approach corresponding to the O’Brien-Fleming boundaries. The monitoring boundaries for moderate evidence (alpha =5%, beta =20%) were based on information size required for moderate evidence (10,508 patients). The monitoring boundaries for strong evidence (alpha =1%, beta =10%) were based on information size required for strong evidence (19,920 patients). Both information size calculations were based on the assumption of a 5% incidence rate in the control group, and a 25% relative risk reduction. The cumulative Z-statistic does not cross above either of the two monitoring boundaries, and thus, there is neither moderate or strong statistical evidence (significance) to support a treatment effect of at least a 25% relative risk reduction.
Figure 3
Figure 3
Trial sequential analysis (TSA) monitoring boundaries for moderate evidence (the downward sloping line) after adding a ‘fake trial’ with 3800 patients. Notes: The monitoring boundaries for moderate evidence (alpha =5%, beta =20%) were constructed using the Lan-DeMets alpha-spending approach corresponding to the O’Brien-Fleming boundaries, and were based on the information size required for moderate evidence (10,508 patients). The information size calculation was based on the assumption of a 5% incidence rate in the control group, and a 25% relative risk reduction. The cumulative Z-statistic crossed the monitoring boundaries for moderate evidence after adding a ‘fake trial’ with an intervention effect estimate of a 25% relative risk reduction, a 5% control group incidence rate and 3800 patients (1900 in each group). This result demonstrates that under the event rate assumptions made for the information size calculations, a new trial would need to include 3800 patients (1900 in each group) to make the meta-analysis conclusive with moderate statistical support.
Figure 4
Figure 4
Trial sequential analysis (TSA) monitoring boundaries for strong evidence (the downward sloping line) after adding a ‘fake trial’ with 9000 patients. Notes: The monitoring boundaries for moderate evidence (alpha =1%, beta =10%) were constructed using the Lan-DeMets alpha-spending approach corresponding to the O’Brien-Fleming boundaries, and were based on the information size required for moderate evidence (19,920 patients). The information size calculation was based on the assumption of a 5% incidence rate in the control group, and a 25% relative risk reduction. The cumulative Z-statistic crossed the monitoring boundaries for moderate evidence after adding a ‘fake trial’ with an intervention effect estimate of a 25% relative risk reduction, a 5% control group incidence rate and 9000 patients (4500 in each group). This result demonstrates that under the event rate assumptions made for the information size calculations, a new trial would need to include 9000 patients (4500 in each group) to make the meta-analysis conclusive with strong statistical support.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

  • Microwave coagulation for liver metastases.
    Storman D, Swierz MJ, Mitus JW, Pedziwiatr M, Liang N, Wolff R, Bala MM. Storman D, et al. Cochrane Database Syst Rev. 2024 Mar 27;3(3):CD010163. doi: 10.1002/14651858.CD010163.pub3. Cochrane Database Syst Rev. 2024. PMID: 38534000 Review.
  • Antibiotics for treatment of leptospirosis.
    Win TZ, Han SM, Edwards T, Maung HT, Brett-Major DM, Smith C, Lee N. Win TZ, et al. Cochrane Database Syst Rev. 2024 Mar 14;3(3):CD014960. doi: 10.1002/14651858.CD014960.pub2. Cochrane Database Syst Rev. 2024. PMID: 38483092 Review.
  • Antibiotic prophylaxis for leptospirosis.
    Win TZ, Perinpanathan T, Mukadi P, Smith C, Edwards T, Han SM, Maung HT, Brett-Major DM, Lee N. Win TZ, et al. Cochrane Database Syst Rev. 2024 Mar 14;3(3):CD014959. doi: 10.1002/14651858.CD014959.pub2. Cochrane Database Syst Rev. 2024. PMID: 38483067 Review.
  • Deep neuromuscular blockade in adults undergoing an abdominal laparoscopic procedure.
    Bijkerk V, Jacobs LM, Albers KI, Gurusamy KS, van Laarhoven CJ, Keijzer C, Warlé MC. Bijkerk V, et al. Cochrane Database Syst Rev. 2024 Jan 30;1(1):CD013197. doi: 10.1002/14651858.CD013197.pub2. Cochrane Database Syst Rev. 2024. PMID: 38288876 Review.
  • Transjugular intrahepatic portosystemic shunts for adults with hepatorenal syndrome.
    Gonzalez-Garay AG, Serralde-Zúñiga AE, Velasco Hidalgo L, Flores García NC, Aguirre-Salgado MI. Gonzalez-Garay AG, et al. Cochrane Database Syst Rev. 2024 Jan 18;1(1):CD011039. doi: 10.1002/14651858.CD011039.pub2. Cochrane Database Syst Rev. 2024. PMID: 38235907
See all "Cited by" articles

References

    1. Guyatt G, Haynes RB, Jaeschke RZ, et al. Users’ Guides to the Medical Literature: XXV. Evidence-based medicine: principles for applying the Users’ Guides to patient care. JAMA. 2002;284:1290–1296. - PubMed
    1. Higgins JP, Green S. Cochrane Handbook for systematic reviews of interventions, version 5.0.0. Oxford: John Wiley & Sons; 2009.
    1. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336:924–926. - PMC - PubMed
    1. LeLorier J, Gregoire G, Benhaddad A, Lapierre J, Derderian F. Discrepancies between meta-analyses and subsequent large randomized, controlled trials. NEJM. 1997;337:536–542. - PubMed
    1. Moses LE, Mosteller F, Buehler JH. Comparing results of large clinical trials to those of meta-analysis. Stat Med. 2002;21:793–800. - PubMed
-