Invited Commentary: Factor Analysis and the Search for Objectivity

Martı́nez, Marı́a Elena; Marshall, James R.; Sechrest, Lee

doi:10.1093/oxfordjournals.aje.a009552

Cited by 295 publications

(254 citation statements)

References 6 publications

Supporting

Mentioning

242

Contrasting

Unclassified

Order By: Relevance

“…However, we used a very conservative eigenvalue criterion to retain only those factors that account for a more substantial proportion of variability in dietary intake, and this may reduce the potential for overfitting the dietary patterns. Although factor analysis involves decisions that can be called subjective or arbitrary, 58 our dietary patterns are generally consistent with patterns in previous studies, especially Healthy Living and Traditional Western.…”

Section: Discussionsupporting

confidence: 80%

Dietary patterns and risk of prostate cancer in Ontario, Canada

Walker

Aronson

King

et al. 2005

Intl Journal of Cancer

View full text Add to dashboard Cite

Dietary patterns reflect combinations of dietary exposures, and here we examine these in relation to prostate cancer risk. In a case-control study, 80 incident primary prostate cancer cases and 334 urology clinic controls were enrolled from 1997 through 1999 in Kingston, Ontario, Canada. Food-frequency questionnaires were completed prior to diagnosis and assessed intake in the 1-year period 2-3 years prior to enrollment. Among controls, dietary intake was used in principal components analyses to identify patterns that were then evaluated with all subjects in relation to prostate cancer risk using unconditional logistic regression, controlling for age. Prostate cancer is one of the most common cancer types in developed countries, with an estimated 20,100 new cases in 2004 in Canada. 1 Higher incidence rates occur in North America, Great Britain, Western Europe and Australia than in Asia, Latin America and Africa. 2 While genetic susceptibility is an established risk factor for prostate cancer, 2-5 the wide international variation in incidence points primarily to an environmental etiology. 6,7 The role of diet has been investigated in several epidemiologic studies, including dietary fat intake, vitamin A, carotenoids, fruits, vegetables, dairy products and soy products. 2,3,[7][8][9][10][11][12][13][14][15][16] The most consistent dietary component associated with prostate cancer risk is fat intake, 2 an association that is apparent despite differences in the quality of the study methods and assessment of dietary intake. [2][3][4]7,8,16 Recent attention has turned to a group of phytochemicals, particularly the isoflavones found in soybeans, 10,15 since Asians, with lower prostate cancer rates, have high consumption of soy products. Hypotheses regarding exposure to cadmium from foods such as shellfish and the risk of prostate cancer have not been substantiated by epidemiologic investigations. 2 Epidemiologic evidence supports an association between dietary factors and the risk of prostate cancer, but there are few consistent findings at a specific level. This is due in part to differences in measures of dietary intake, lack of variability in nutrient intakes among the population being studied and lack of control in the analysis for several possible confounders, including energy intake and lifestyle factors such as physical activity and smoking. 2,17,18 Further, potential associations between diet and prostate cancer risk may be concealed or lost when individual food items or nutrients are analyzed for their association with risk, as opposed to looking at groupings of habitually consumed foods. The metabolic and physiologic effects of combinations of foods on total risk are also unaccounted for when the conventional method of nutrient or individual food item analysis is used. 19 A broader consideration of patterns of food use may provide a more comprehensive assessment of the role of dietary factors on prostate cancer risk, given that foods are generally eaten in combination and together may have a greater impact on...

show abstract

Section: Discussionsupporting

confidence: 80%

Dietary patterns and risk of prostate cancer in Ontario, Canada

Walker

Aronson

King

et al. 2005

Intl Journal of Cancer

View full text Add to dashboard Cite

show abstract

“…Finally, the limitations and subjective nature of factor analysis technique are widely acknowledged. 15,24,46,47 However, to minimize subjectivity, food groupings used were similar to those reported by others and the selection of the factor solution was done after evaluating scree plot and eigenvalues.…”

Section: Discussionmentioning

confidence: 99%

Associations between dietary patterns and obesity phenotypes

et al. 2009

View full text Add to dashboard Cite

Objective: To examine whether dietary patterns are associated with obesity phenotypes. Design: Cross-sectional study. Subjects: We recruited 664 participants aged between 18 and 55 years. Dietary data were collected from a food frequency questionnaire. A factor analysis was performed to derive dietary patterns. Body mass index (BMI), weight and waist girth were recorded using standard procedures. Fat mass and fat-free mass were assessed by electrical bioimpedance. Obesity was defined as having a BMIX30 kg m À2 and a positive FHO (FHO þ ) as having at least one obese first-degree relative. Results: Two dietary patterns were identified; Western and Prudent. The Western pattern was mainly characterized by a higher consumption of refined grains, French fries, red meats, condiments, processed meats and regular soft drinks whereas the Prudent pattern was mainly characterized by a higher consumption of non-hydrogenated fat, vegetables, eggs and fish and seafood. Subjects in the top tertile of the Western pattern had higher BMI, weight, waist girth, waist-to-hip ratio and fat mass than those in the lower tertile. In contrast, subjects in the top tertile of the Prudent pattern had lower BMI, weight, waist girth, fat mass, HDL-cholesterol levels, and lower triglyceride levels than those in the lowest tertile. Individuals in the upper tertile of the Western pattern were more likely to be obese (obesity was defined as having a BMIX30 kg m À2 ) (OR ¼ 1.82, 95% CI 1.16-2.87) whereas those in the upper tertile of the Prudent pattern were less likely to be obese (OR ¼ 0.62, 95% CI 0.40-0.96). These latter significant associations were only observed among those with FHO þ . No such association was observed among FHOÀ individuals. Conclusion: Individuals having a high score of Western pattern were more likely to be obese and those having a high score of the Prudent pattern were less likely to be obese, and this is particularly among individuals with an FHO þ .

show abstract

“…However, the complexity makes it also more prone to subjective interpretations (34) . We wanted to evaluate whether the dietary patterns carried any additional relationship with weight gain rate that was beyond detection by nutrient-level analysis.…”

Section: Discussionmentioning

confidence: 99%

Unhealthy dietary patterns are associated with weight gain during pregnancy among Finnish women

Uusitalo

Arkkola

Ovaskainen

et al. 2009

Public Health Nutr.

View full text Add to dashboard Cite

Objective: To study whether the dietary patterns of Finnish pregnant women are associated with their weight gain rate during pregnancy. Design: A validated 181-item FFQ was applied retrospectively to assess the diet during the eighth month of pregnancy, and maternal height and maternal weight at first and last antenatal visits were recalled. Information on sociodemographic characteristics, parity and smoking of the pregnant women was obtained by a structured questionnaire and from the Finnish Birth Registry. Principal components analysis was used to identify dietary patterns that described the diet of pregnant women based on their food consumption profile. Setting: Finland. Subjects: Subjects consisted of 3360 women who had newly delivered in 1997-2002 and whose baby carried human leucocyte antigen-conferred susceptibility to type 1 diabetes in two university hospital regions, Oulu and Tampere, in Finland. Results: Out of seven dietary patterns identified, the 'fast food' pattern was positively associated (b 5 0?010, SE 5 0?003, P 5 0?004) and the 'alcohol and butter' pattern was inversely associated (b 5 -0?010, SE 5 0?003, P , 0?0001) with weight gain rate (kg/week) during pregnancy after adjusting for potential dietary, perinatal and sociodemographic confounding factors. Both of the dietary pattern associations demonstrated dose dependency. Conclusions: Pregnant women should be guided to have a well-planned, balanced, healthy diet during pregnancy in order to avoid rapid gestational weight gain. The association between diet, health and maternal weight gain of the women who consumed alcohol during pregnancy should be studied further.

show abstract

Invited Commentary: Factor Analysis and the Search for Objectivity

Cited by 295 publications

References 6 publications

Dietary patterns and risk of prostate cancer in Ontario, Canada

Dietary patterns and risk of prostate cancer in Ontario, Canada

Associations between dietary patterns and obesity phenotypes

Unhealthy dietary patterns are associated with weight gain during pregnancy among Finnish women

Contact Info

Product

Resources

About