Çok merkezli sağlık araştırmalarında tasarım etkisinin tahmini ve örnek büyüklüğüne etkisi

Amaç: Bu makalenin amacı, tasarım etkisi (Deff) değerinden yararlanarak, çok merkezli araştırmalarda farklı sınıf içi korelasyon katsayıları için örneklem büyüklüğünü incelemektir. Yöntem: Çok merkezli çalışmalar çok sayıda metodolojik ve ekonomik avantaj sağlar. Kümelenmiş / gruplanmış bir yapının söz konusu olduğu durumlarda küme ilişkili veriler ile karşılaşılmaktadır. Bu tür veriler, bireylerin çok farklı şekillerde gruplanabilir olmasından dolayı, çoğunlukla sosyal, davranış ve sağlık bilimlerinde ortaya çıkmaktadır. Basit rasgele örnekleme yöntemine göre kompleks bir tasarımla üretilen tahminlerin hassaslığındaki fark, tasarım etkisi olarak bilinmektedir. Tasarım etkisi bir araştırma istatistiğidir. Büyük ölçekli örneklem araştırmalarında, çıkarımlar genellikle araştırma yapılan örneklemin rasgeleliği ilkesine dayanır. Böyle bir yaklaşımla, rasgeleliğin yalnızca örneklemin oluşturulmasındaki olasılık mekanizmasından kaynaklandığı varsayılır. Bulgular: Tasarım etkisinin iyi bir tahmini kümelenmenin söz konusu olduğu araştırmalarda en uygun örneklem büyüklüğünü hesaplamak için kritik önem taşır. Kümeleme, gerçek kitle varyansını olduğundan daha düşük tahmin eder ve bu eğer doğru tahmin edilmiş ise, aynı büyüklükteki basit rasgele örneklemden elde edilecek standart hatalardan büyük olan standart hatalara yansır. Sonuç: Tasarım etkisi ölçülen göstergeye göre kümeler arasındaki heterojenliği hesaplamak için bir "düzeltme faktörü" dür. Unutulmamalıdır ki, Deff bir çarpım faktördür, bu nedenle bir araştırmada Deff değerinin 2 olarak hesaplanması, araştırmada dikkate alınacak örneklem büyüklüğünün iki kat daha fazla olması demektir.

Anahtar Kelimeler:

Çok merkezli araştırma, küme tabanlı araştırma, örneklem büyüklüğü, araştırma tasarımı, tasarım etkisi

Estimation of design effect in multicentre health research and its effect on sample size

Objective: The purpose of this article is to investigate the sample size of different intraclass correlation coefficients in multicenter studies, taking advantage of the Deff value. Method: Multicenter studies bring numerous methodological and economic advantages. Cluster-correlated data arises when there is a clustered/grouped structure to the data. Data of this kind frequently arises in the social, behavioral, and health sciences since individuals can be grouped in many different ways. The difference in the precision of the estimates produced by a complex design relative to a simple random sample is known as the design effect (Deff). The Deff is a survey statistic. In large-scale sample surveys, inferences are usually based on the standard randomization principle of survey sampling. Under such an approach, the responses are treated as fixed, and the randomness is assumed to come solely from the probability mechanism that generates the sample. Results: A good estimate of the Deff is critical for calculating the most efficient sample size for cluster surveys. Clustering, underestimates true population variance and this is reflected in standard errors that are larger, if correctly estimated, than those that would have been obtained from a simple random sample of the same size. Conclusions: The Deff is a “correction factor” to account for the heterogeneity between clusters with regard to the measured indicator. It should be remembered, Deff is a multiplying factor, so if the value of Deff in the survey is calculated as 2, this means that the sample size to be taken into account in the survey is twice as large.

Keywords:

Multicenter research, cluster based survey, sample size, research design, design effect,

PDF

___

Anello C, O’Neill RT, Dubey S. Multicentre trials: a US regulatory perspective. Stat Methods Med Res 2005; 14(3): 303-318.
Vierron E, Giraudeau B. Sample size calculation for multicenter randomized trial: taking the centre effect into account. Contemp Clin Trials 2007; 28(4): 451-458.
Localio AR, Berlin JA, Ten Have TR, Kimmel SE. Adjustments for center in multicenter studies: an overwiev. Ann Inter Med 2001; 135(2): 112-123.
Bland JM. Cluster randomised trials in the medical literature: two bibliometric surveys. BMC Med Res Methodol 2004; 4: 21.
Perera R, Glasziou P. A simple method to correct for the design effect in systematic reviews of trials using paired dichotomous data. J Clin Epidemiol 2007; 60(9): 975-978.
Gambino JG. Design effect caveats. Am Stat 2009; 63(2): 141-146.
Kish L. (1965) Survey Sampling. John Wiley and Sons, Inc. London.
Liu J, Iannacchione V, Byron M. Decomposing design effects for stratified sampling. In: Proceedings of the American Statistical Association’s Section on Survey Section on Survey Research Methods – JSM 2012 Research Methods, 2012.
Kaiser R, Woodruff BA, Bilukha O, Spiegel PB, Salama P. Using design effects from previous cluster surveys to guide sample size calculation in emergency settings. Disasters 2006; 30(2): 199-211.
Bell BA, Onwuegbuzie AJ, Ferron JM, Jiao QG, Hibbard ST, Kromrey JD. Use of design effects and sample weights in complex health survey data: A review of published articles using data from 3 commonly used adolescent health surveys. Am J Public Health 2012; 102(7): 1399-1405.
Vierron E, Giraudeau B. Design effect in multicenter studies: gain or loss of power? BMC Med Res Methodol 2009; 9(39). doi:10.1186/1471-2288-9-39.
Walker DA, Young DY. Example of the impact of weights and design effects on contingency tables and Chi-Square analysis. JMASM 2003; 2(2): 425-432.
Kerry SM, Bland JM. The intracluster correlation coefficient in cluster randomisation. Brit Med J 1998; 316(7142): 1455.
Kalton G, Brick JM, Le T. Estimating components of design effects for use in sample design. In: Household Sample Surveys in Developing and Transition Countries. Department of Economic and Social Affairs Statistics Division Studies in Methods Series F No. 96. United Nations New York; 2005. p. 98-121.
Chatrchi G, Brisebois F. Survey weighting adjustments and the design effect: A case study. In: Proceedings of the American Statistical Association’s Section on Survey Section on Survey Research Methods – JSM 2015 Research Methods, 2015.
Kish L. Methods for design effects. J Off Stat 1995; 11(1): 55-77.
Johnson DR, Elliott LA. Sampling design effects: Do they affect the analyses of data from the national survey of families and households? J Marriage Fam 1998; 60(4): 993-1001.
Williams B, Gopi PG, Borgdorff MW, Yamada N, Dye C. The design effect and cluster samples: Optimising tuberculosis prevalence surveys. Int J Tuberc Lung D 2008; 12(10): 1110–1115.
Mickey RM, Goodwin GD, Costanza MC. Estimation of the design effect in community intervention studies. Stat Med 1991; 10(1): 53-64.
Henry KA, Richard Valliant R. A design effect measure for calibration weighting in single-stage samples. Surv Methodol 2015; 41(2): 315-331.
Wejnert C, Pham H, Krishna N, Le B, DiNenno E. Estimating design effect and calculating sample size for respondent-driven sampling studies of injection drug users in the United States. AIDS Behav 2012; 16(4): 797-806.
Coupland C, DiGuiseppi C. The design and use of cluster randomised controlled trials in evaluating injury prevention interventions: part 2. Design effect, sample size calculations and methods for analysis. Inj Prev 2010; 16(2): 132-136.
Masood M, Reidpath DD. Intraclass correlation and design effect in BMI, physical activity and diet: a cross-sectional study of 56 countries. BMJ Open 2016; 6: e008173.
Janjua NZ, Khan MI, Clemens JD. Estimates of intraclass correlation coefficient and design effect for surveys and cluster randomized trials on injection use in Pakistan and developing countries. Trop Med Int Health 2006; 11(12): 1832-1840.
Katz J, Zeger SL. Estimation of design effects in cluster surveys. Ann Epidemiol 1994; 4(4): 295-301.
Hulland EN, Blanton CJ, Leidman EZ, Bilukha OO. Parameters associated with design effect of child anthropometry indicators in small‑scale field surveys. Emerg Themes Epidemiol 2016; 13: 13.
Salganik MJ. Variance estimation, design effects, and sample size calculations for respondent-driven sampling. B New York Acad Med 2006; 83(7): 98-112.
Shojania KG, Ranji SR, Shaw LK, Charo LN, Lai JC, Rushakoff RJ, McDonald KM, Owens DK. Diabetes Mellitus Care. Vol. 2 of: Shojania KG, McDonald KM, Wachter RM, Owens DK. Closing The Quality Gap: A Critical Analysis of Quality Improvement Strategies. Technical Review 9 (Contract No. 290-02-0017 to the Stanford University–UCSF Evidence-based Practice Center). AHRQ Publication No. 04-0051-2. Rockville, MD: Agency for Healthcare Research and Quality. 2004. p. 168.
Jabkowski P. How (not) to estimate the design effect of a complex sampling scheme: A case study of the Polish section of the European Social Survey, round 5. Research and Methods, 2013; 22(1): 55–77.
Unverzagt S, Prondzinsky R, Peinemann F. Single-center trials tend to provide larger treatment effects than multicenter trials: a systematic review. J Clin Epidemiol 2013; 66(11): 1271-1280.
Sprague S, Matta JM, MD, Bhandari M. Multicenter collaboration in observational research: improving generalizability and efficiency. J Bone Joint Surg 2009; 91(3): 80-86.
Messerer D, PorzsoIth F, Hasford J, Neiß A. Vorteile und probleme multizentrischer therapiestudien am Beispiel einer studie zur behandlung des metastasierenden nierenzellkarzinoms mit rekombinantem Interferon-Alpha-ZC1. Onkologie, 1987; 10(1): 43-49.
Greene SM, Geiger AM. A review finds that multicenter studies face substantial challenges but strategies exist to achieve Institutional Review Board approval. J Clin Epidemiol 2006; 59(8): 784–790.
Lin Z. The number of centers in a multicenter clinical study effects on statistical power. Drug Inf J 2000; 34(2): 379-386.
Ruvuna F. Unequal center sizes, sample size, and power in multicenter clinical trials. Drug Inf J 2004; 38(4): 387-394.