Abstract
Complementary alternative medicine, especially dietary supplements (DS), has gained increasing popularity for weight loss due to its availability without prescription, price, and ease of use. Besides weight loss, there are various perceived, potential benefits linked to DS use. However, health consumers with limited health literacy may not adequately know the benefits and risk of overdose for DS. In this project, we aim to gain a better understanding of the use of DS products among obese people as well as the perceived benefits of these products. We identified obese adults after combining the National Health and Nutrition Examination Survey data collected from 2003 to 2014. We found that there is a knowledge gap between the reported benefits of major DS by obese adults and the existing DS knowledge base and label database. This gap may inform the design of patient education material on DS usage in the future.
Introduction
Obesity, a complex chronic medical condition having multifactorial etiology, is on the rise not only in the US but also around the globe.1 In the US, the prevalence of obesity among population 18 years or older has increased from 33.7% in 2007-2008 to 39.6% in 2015-2016, whereas prevalence of severe obesity in adults has increased from 5.7% in 2007-2008 to 7.7% in 2015-2016.2 The resultant, global epidemic is considered to be a consequence of an imbalance between reduced energy expenditure and increased energy intake.3 Low socioeconomic status such as lower income and lower level of education is associated with a greater risk of obesity.4 It was estimated that an obese person in the US incurs an average of over $1,400 more in annual medical expenses, accumulating to approximately $147 billion in medical expenses spent per year in the US.5 Moreover, obesity has been reported to contribute to about 100,000 – 400,000 deaths per year in the US.6
Obesity in adults is defined as a body mass index (BMI, i.e., weight in kilograms divided by height in meters squared) of ≥ 30 kg/m2or more. Obesity is often represented in terms of three categories based on BMI values, i.e., Class 1 (BMI=30 to 35 kg/m2), Class 2 (BMI= 35 to 40 kg/m2), and Class 3 (BMI ≥ 40 kg/m2). Class 3 obesity is sometimes categorized as “extreme” or “severe” obesity. Obesity is often a predecessor of chronic, more serious medical issues (e.g., Type 2 diabetes mellitus, hyperlipidemia, ischemic heart disease, stroke),7 and/or mental health conditions (e.g., depression, anxiety, eating disorders),8 in addition to functional limitations ensuing in poor quality of life and high mortality rates.9 Apart from public health concerns, the high prevalence of adult obesity also poses a substantial financial and economic burden resulting from extravagant medical care costs.10
Currently, there are a number of available options out there to simply maintain a healthy weight, or to actually treat people with obesity or those who are at a high risk of weight-related comorbidities. This ranges from lifestyle changes (e.g., exercise, diet modifications) to more aggressive treatments, i.e., pharmacological (e.g., Phentermine, Orlistat, Lorcaserin, Naltrexone/Bupropion) and/or invasive bariatric surgeries (e.g., gastric banding, gastric bypass, Roux-en-Y gastric bypass).9 However, the vast majority of treatment around lifestyle changes often fail due to noncompliance resulting from various factors, e.g., poor motivation, lack of time, gaps in knowledge/awareness, lack of strong commitment to seeing actual results.11 On the other hand, most of the pharmacological and surgical procedures are associated with a substantial amount of health risk and at the cost of large dollar amounts.9 Hence, people often are looking for alternative therapies (e.g., supplements, acupuncture, non-invasive body-contouring) which are effective, quick-acting with minimal side effects, easily available, and relatively cheap. Many patients who are overweight and obese consider using dietary supplements for weight loss.
The popularity of complementary alternative medicine (CAM), especially the use of dietary supplements (DSs) for weight management, has gained much popularity. Aside from the health benefits resulting from weight loss, there are various other reasons for people to turn to DSs for losing weight and/or maintaining a healthy weight, e.g., resulting frustration from failures in previous weight loss attempts following strict diet and exercise regimens, DSs being falsely considered as “magic bullets” that are “natural and safe”, easy availability without a prescription, easy to take, less expensive, bypassing the physician’s office visit fee, inflated advertising claims.12 Some of the commonly used DSs classified according to their proposed mechanism of actions are DSs: stimulating energy expenditure (e.g., Ephedra, Bitter orange), modulating carbohydrate metabolism (e.g., Chromium, Ginseng), increasing satiety (e.g., Guar gum, Psyllium), increasing fat oxidation or decreasing fat synthesis (e.g., Green tea, Licorice), blocking absorption of fat (e.g., Chitosan), eliminating excessive body water (e.g., Dandelion, Cascara), enhancing mood (e.g., St. John’s wart), and others (e.g., Laminaria, Apple cider vinegar).12
Interestingly, earlier studies have revealed that the use of DSs is more commonly preventive with an aim to maintain and improve overall health, rather than being therapeutic in order to treat obesity.13 According to the National Health and Nutrition Examination Survey (NHANES) 2003–2006, a nationally representative, cross-sectional survey, obese respondents reported relatively less DS use (48%) than those categorized as overweight (57%) or normal weight (56%).13 Hence, DS users, as compared to non-users, are significantly more likely to have better dietary patterns, exercise regularly, maintain a healthy weight, and avoid tobacco products.13 It is also known that DS consumers with limited health literacy may not adequately know the benefits and risk of overdose for DSs.14 Thus, it is critical to create a DS profile for general health consumers to inform the design of patient education material for DSs.
In this project, we aim to gain a better understanding of the use of DS products among obese people as well as their perceived benefits of these products. We used the combined NHANES data from 2003-2014 to answer three research questions (RQ):
RQ1: What are the perceived benefits of DS use for patients with obesity?
RQ2: Are there associations between patients’ socioeconomic status and demographic information with DS use?
RQ3: Is there a knowledge gap between the benefits for DS for patients with obesity and existing DS knowledge bases?
Methods
Data Source
NHANES is a continuous cross-sectional health survey conducted by the National Center for Health Statistics of CDC.15 It evaluates a stratified multistage probability sample of the non-institutionalized population of the United States. The survey samples are first interviewed at home, followed by a physical and laboratory test in a mobile examination center. Its rigorous quality control standards ensure national representativeness and high-quality data collection. NHANES data have facilitated various public health16 and biomedical informatics research.17-19
Data Preparation
We first extracted the demographic, examination, and dietary data from NHANES for survey years 2003 – 2014 (6 survey cycles). To strengthen the analytical power of the study, survey data from multiple survey cycles were combined for the following analyses. Inclusion criteria included: (1) DS use information, and (2) age ≥ 18. We split this cohort into a control and obese group based on BMI, with the obese group including all participants having BMI ≥ 30 kg/m2. From NHANES data, we removed 1937 participants (5.4%) with no BMI values. We also removed 27 participants with no values for taking DS or not. The “wtint2yr” variable (2-year sample weight for interviewed participants) is a sample weight assigned to each sample person by NHANES to match U.S. census population totals. It represents the number of non-institutional people in the US that a survey participant can represent. According to NHANES analytical guideline1, when appropriately combining all 6 survey cycles, we divided “wtint2yr” by 6 to construct new sample weights before the analyses. Note that the one person only appeared in one survey cycle. After applying sample weights, the control group includes the remaining 21,997 respondents with a total sample weight of 140,431,403 and the obese group includes 11,954 respondents with a sample weight of 73,697,181.
DS use was pulled for this cohort. Total and individual DS use was available for all survey cycles although detailed data was inconsistent for years 2003-2004 and 2007-2008. These inconsistencies caused minor issues with data processing but not with the data validity. DSs used were grouped into types based on product information. Detailed information on reasons for DS use was available for survey cycles from 2007-2014.
Data Analysis
Basic Characteristics: We first created a profile of the cohort with respect to gender, age, race, and household income. On the more detailed data from 2007-2014, we assessed the major perceived benefits of the DSs used by the cohort, stratified by specific DS type.
Comparing the Reported DS Use with Existing Knowledge Bases: We analyzed the individual reasons given by survey participants for years 2007-2014 related to DS use. In addition, we also investigated if the information provided in the existing knowledge base aligns with the reported use of a particular DS. In our previous study, a qualitative evaluation was performed (RR- a coauthor, physician and health informaticist) and compared across five selected databases for presences of essential data elements after cross-checking them against a preliminary, standardized set of data elements.20 It was learned that Natural Medicine Comprehensive Database (NMCD)21 was found to be the most comprehensive of all the resources, providing DS information that is reliable, clinically relevant data and evidence-based, monitored and updated regularly. Hence, in this study, we compared the reported benefits of obese adults in NHANES with NMCD aiming at identifying the knowledge gap.
Correlation Analysis: We used Pearson correlation coefficient to assess the correlation between numeric variables (e.g., age and the number of DSs taken). We also used multivariate logistic regression to evaluate the impact of basic characteristics on DS use between control and obese groups. In addition, we attempted to use analysis of covariance (ANCOVA) for the correlation analysis between a numeric and categorical variable (e.g., household income and DS use). ANCOVA is a method that is subject to a number of assumptions including: 1) linear relationship between dependent and independent variables, 2) independent homogeneous normally distributed error, 3) homogeneity of regression slopes between groups. Unfortunately, these assumptions were not met in our case, thus ANCOVA results are not reportable in this study.
Results
The average BMI for males in the obese group was 35 kg/m2. The average BMI for females in the obese group was 36.53 kg/m2. Table 1 shows the basic characteristics of the study population. Out of the 33,951 survey participants, 51.26% (weighted frequency, WF, is 109,766,046) self-report taking at least one DS. The control group accounts for 74,458,068 of this and the obese group accounts for 35,307,978. The participants in survey cycles 2007-2014, they were asked if ‘they took the product on their own‘ (Self) or ‘doctor or health provider told me to’ (HCP). For this group of survey participants (taking WF 170,115,087 DSs in total), 74.12% were taken due to their personal will and the remaining 25.88% were taken due to the advice of a health care practitioner. When looking at the control the WF are as follows “Self” – 88,231,094 (76.44%) and “HCP” – 27,200,815 (23.56%). The obese group is 37,855,117 (69.23%) and 16,828,061 (30.77%) respectively. Wald Chi-square tests showed statistically significant differences between the control and obese groups with respect to gender, age group, race, and household income (P<.0001).
Table 1.
Basic characteristics of the study population
| Variable | Overall | 0 DS | 1 or more DS | |||
|---|---|---|---|---|---|---|
| Control | Obese | Control | Obese | Control | Obese | |
| Gender | WF (%) | WF (%) | WF (%) | WF (%) | WF (%) | WF (%) |
| Male | 69,398,631 (49.42%) | 33,945,341 (46.06%) | 37,481,223 (54.01%) | 19,739,505 (58.15%) | 31,917,408 (45.99%) | 14,205,836 (41.85%) |
| Female | 71,032,772 (50.58%) | 39,751,841 (53.94%) | 28,492,112 (40.11%) | 18,649,699 (46.92%) | 42,540,660 (59.89%) | 21,102,142 (53.08%) |
| Age | ||||||
| 18-24 | 21,130,392 (15.05%) | 6,556,637 (8.90%) | 14,144,116 (66.94%) | 4,676,915 (71.33%) | 6,986,276 (33.06%) | 1,879,722 (28.67%) |
| 25-34 | 25,791,429 (18.37%) | 11,813,620 (16.03%) | 14,561,110 (56.46%) | 7,778,604 (65.84%) | 11,230,319 (43.54%) | 4,035,016 (34.16%) |
| 35-44 | 25,153,315 (17.91%) | 14,941,874 (20.27%) | 12,795,306 (50.87%) | 9,218,501 (61.70%) | 12,358,010 (49.13%) | 5,723,372 (38.3%) |
| 45-54 | 26,139,764 (18.61%) | 15,787,763 (21.42%) | 11,529,146 (44.11%) | 8,009,651 (50.73%) | 14,610,617 (55.89%) | 7,778,111 (49.27%) |
| 55-64 | 18,781,403 (13.37%) | 12,794,888 (17.36%) | 6,469,237 (34.44%) | 4,844,214 (37.86%) | 12,312,166 (65.56%) | 7,950,674 (62.14%) |
| 65-74 | 12,680,140 (9.03%) | 7,809,087 (10.60%) | 3,670,919 (28.95%) | 2,598,196 (33.27%) | 9,009,221 (71.05%) | 5,210,891 (66.73%) |
| 75 and over | 10,754,960 (7.66%) | 3,993,312 (5.42%) | 2,803,501 (26.07%) | 1,263,121 (31.63%) | 7,951,459 (73.93%) | 2,730,191 (68.37%) |
| Race | ||||||
| Mexican American | 10,937,873 (7.79%) | 7,093,000 (9.62%) | 7,503,518 (68.60%) | 4,865,723 (68.60%) | 3,434,356 (31.40%) | 2,227,277 (31.40%) |
| Other Hispanic | 6,838,482 (4.87%) | 3,634,863 (4.93%) | 4,059,906 (59.37%) | 2,222,149 (61.13%) | 2,778,576 (40.63%) | 1,412,714 (38.87%) |
| Non-Hispanic White | 98,016,824 (69.80%) | 48,756,475 (66.16%) | 40,424,225 (41.24%) | 22,858,918 (46.88%) | 57,592,599 (58.76%) | 25,897,557 (53.12%) |
| Non-Hispanic Black | 13,389,691 (9.53%) | 11,304,379 (15.34%) | 8,518,257 (63.62%) | 6,931,321 (61.32%) | 4,871,434 (36.38%) | 4,373,057 (38.68%) |
| Other Race or Multi- | 11,248,533 | 2,908,465 | 5,467,429 | 1,511,093 | 5,781,104 | 1,397,372 |
| Racial | (8.01%) | (3.95%) | (48.61%) | (51.96%) | (51.39%) | (48.04%) |
| Household Income | ||||||
| $0 to $4,999 | 2,185,115 (1.56%) | 1,040,625 (1.41%) | 1,334,103 (61.05%) | 703,261 (67.58%) | 851,012 (38.95%) | 337,364 (32.42%) |
| $5,000 to $9,999 | 3,642,234 (2.59%) | 2,363,568 (3.21%) | 21,68,371 (59.53%) | 1,433,092 (60.63%) | 1,473,863 (40.47%) | 930,476 (39.37%) |
| $10,000 to $14,999 | 7,013,669 (4.99%) | 4,088,184 (5.55%) | 3,977,111 (56.71%) | 2,358,676 (57.70%) | 3,036,558 (43.29%) | 1,729,508 (42.30%) |
| $15,000 to $19,999 | 6,979,855 (4.97%) | 3,925,689 (5.33%) | 3,988,366 (57.14%) | 2,296,441 (58.50%) | 2,991,489 (42.86%) | 1,629,248 (41.50%) |
| $20,000 to $24,999 | 8,526,874 (6.07%) | 4,840,260 (6.57%) | 4,577,403 (53.68%) | 2,797,069 (57.79%) | 3,949,471 (46.32%) | 2,043,190 (42.21%) |
| $25,000 to $34,999 | 13,829,794 (9.85%) | 8,092,533 (10.98%) | 7,328,211 (52.99%) | 4,536,612 (56.06%) | 6,501,583 (47.01%) | 3,555,920 (43.94%) |
| $35,000 to $44,999 | 12,077,157 (8.60%) | 7,681,006 (10.42%) | 5,822,324 (48.21%) | 3,810,995 (49.62%) | 6,254,834 (51.79%) | 3,870,010 (50.38%) |
| $45,000 to $54,999 | 11,950,643 (8.51%) | 6,404,692 (8.69%) | 6,509,041 (45.53%) | 3,325,334 (51.92%) | 5,441,602 (54.47%) | 3,079,358 (48.08%) |
| $55,000 to $64,999 | 9,129,822 (6.50%) | 5,057,820 (6.86%) | 4,297,419 (47.07%) | 2,479,668 (49.03%) | 4,832,404 (52.93%) | 2,578,153 (50.97%) |
| $65,000 to $74,999 | 7,942,767 (5.66%) | 5,039,288 (6.84%) | 3,494,125 (43.99%) | 2,541,089 (50.43%) | 4,448,642 (56.01%) | 2,498,199 (49.57%) |
| $75,000 and over | 46,857,423 (33.37%) | 20,313,403 (27.56%) | 17,870,026 (38.14%) | 9,451,482 (46.53%) | 28,987,397 (61.86%) | 10,861,921 (53.47%) |
| No answer | 10,296,050 (7.33%) | 4,850,113 (6.58%) | 341,156 (55.11%) | 2,655,484 (54.75%) | 276,400 (44.89%) | 2,194,629 (45.25%) |
We looked at the individual reasons given by survey participants for years 2007-2014 related to DS use. Participants were given a list of options to choose from for each DS. They could choose one or more options from the given list. Table 2 shows the top five reasons. In addition, we also looked at DS use for two additional reasons that were believed to be applicable to the population we are studying. These included ‘For Weight Loss’ and ‘To Maintain Blood Sugar/Diabetes’. Reasons for DS use were matched to specific DS type. Table 2 shows the type of DS with the highest weighted frequency and percent of total response for each reason. The most frequent DS types for the obese group and the control group are the same across all the reasons.
Table 2.
Reason for DS use matched to type of DS
| Reason | Group | Most frequent DS type | Weighted Frequency of responses for this | % of total responsesfor this reason |
|---|---|---|---|---|
| General Overall Healtha | Control | MVMMc | 26,453,115 | 42.38% |
| Obese | MVMMc | 12,189,835 | 42.58% | |
| Bone and Joint Healthb | Control | Calcium / Bone/ Joint | 12,815,331 | 63.61% |
| Obese | Calcium / Bone/ Joint | 5,192,835 | 57.23% | |
| To Supplement Diet/Food Not Enough | Control | MVMMc | 7,851,917 | 54.41% |
| Obese | MVMMc | 3,335,217 | 50.64% | |
| Heart Health/Cholesterol | Control | Omega-3 | 4,770,226 | 48.84% |
| Obese | Omega-3 | 2,814,685 | 53.79% | |
| To Get More Energy | Control | MVMMc | 3,613,966 | 42.51% |
| Obese | MVMMc | 1,999,447 | 44.60% | |
| For Weight Loss | Control | MVMMc | 356,359 | 32.36% |
| Obese | MVMMc | 402,195 | 30.57% | |
| To Maintain Blood Sugar/Diabetes | Control | MVMMc | 283,031 | 36.00% |
| Obese | MVMMc | 251,621 | 29.10% |
a : Includes: To prevent health problems, to improve my overall health, to maintain health/to stay healthy, and to prevent colds/boost immune system.
b : Includes: For healthy joints/arthritis and for bone health/build strong bones/osteoporosis.
c : Multivitamins/multiminerals
The top five types based on the weighted frequency of DS are shown in Table 3.
Table 3.
Top five types of DS
| Dietary Supplement | Control (Weighted Frequency) / (%) | Obese (Weighted Frequency) / (%) |
|---|---|---|
| MVMM | 57,398,221 /32.25% | 25,997,389 /32.03% |
| Calcium / Bone / Joint | 25,353,315 /14.25% | 10,730,464 /13.22% |
| Vitamin B / B-Complex | 15,114,941/ 8.49% | 7,646,862/ 9.42% |
| Omega-3 | 13,742,677/ 7.72% | 6,678,531/ 8.23% |
| Botanicala | 13,059,133/ 7.34% | 6,009,226/ 7.40% |
a: DS classified as a botanical if it is part of a plant, tree, shrub, herb, etc.
Table 4 shows the top five DS types compared to the seven reasons we focused on. In addition, we also investigated if the information provided in the existing knowledge base aligns with the reported use of a particular DS. We found consistency between the reported use of a particular DS (as extracted from NHANES data) for conditions like general overall health, bone and joint health, supplementing food and diets, and heart health/cholesterol, and its use/effectiveness provided in the existing knowledge base (i.e., NMCD). In fact, we found additional useful information about the primary use of a particular DS in addition to its other uses, e.g., use of calcium mainly for bone and joint health other than to improve general health.
Table 4.
DS types and comparison of reported benefits for DS use compared to known knowledge bases
| MVMM | Calcium / Bone / Joint | Vitamin B / B-Complex | Omega-3 | Botanical | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Reason | Group | Weighted Frequency (%) | Supported by the KB? | Weighted Frequency (%) | Supported by the KB? | Weighted Frequency (%) | Supported by the KB? | Weighted Frequency (%) | Supported by the KB? | Weighted Frequency (%) | Supported by the KB? |
| General Overall Healtha | Control | 26,453,115 (42.4%) | Primarily | 4,013,511 (6.4%) | Yes | 4,436,521 (7.1%) | Primarily | 5,613,467 (9.0%) | Yes | 3,319,571 (5.3%) | Yes |
| Obese | 12,189,835 (42.6%) | 1,661,002 (5.8%) | 2,174,237 (7.6%) | 2,412,219 (8.4%) | 1,653,574 (5.8%) | ||||||
| Bone and Joint Healthb | Control | 1,631,498 (8.1%) | Yes | 12,815,331 (63.6%) | Primarily | 221,802 (1.1%) | Yes | 1,041,431 (5.2%) | Yes | 446,372 (2.2%) | Yes |
| Obese | 853,981 (9.4%) | 5,192,835 (57.2%) | 165,819 (1.8%) | 554,147 (6.1%) | 156,958 (1.7%) | ||||||
| To Supplement Diet / Food Not Enough | Control | 7,851,917 (54.4%) | Primarily | 946,322 (6.6%) | Yes | 1,203,612 (8.3%) | Primarily | 725,957 (5.0%) | Yes | 432,633 (3.0%) | Yes |
| Obese | 3,335,217 (50.6%) | 482,093 (7.3%) | 384,737 (5.8%) | 482,233 (7.3%) | 106,195 (1.6%) | ||||||
| Heart Health / Cholesterol | Control | 784,580 (8.0%) | Yes | 179,802 (1.8%) | Conflicting | 734,043 (7.5%) | Yes | 4,770,226 (48.8%) | Primarily | 759,119 (7.8%) | Yes |
| Obese | 453,180 (8.7%) | 61,349 (1.2%) | 500,191 (9.6%) | 2,814,685 (53.8%) | 300,371 (5.7%) | ||||||
| To Get More Energy | Control | 3,613,966 (42.5%) | Yes | 183,743 (2.2%) | No | 2,406,102 (28.3%) | Yes | 236,695 (2.8%) | No | 546,402 (6.4%) | Yes |
| Obese | 1,999,447 (44.6%) | 68,179 (1.5%) | 1,246,553 (27.8%) | 93,423 (2.1%) | 295,225 (6.6%) | ||||||
| For Weight Loss | Control | 356,359 (32.4%) | Noc | 4,170 (0.4%) | Noc | 62,440 (5.7%) | Noc | 48,707 (4.4%) | Yes | 205,332 (18.6%) | Yes |
| Obese | 402,195 (30.6%) | 53,291 (4.1%) | 87,574 (6.7%) | 74,769 (5.7%) | 278,418 (21.2%) | ||||||
| To Maintain Blood Sugar | Control | 283,031 (36.0%) | Noc | 29,274 (3.7%) | Insufficient | 54,507 (6.9%) | Noc | 39,871 (5.1%) | Yes | 137,073 (17.4%) | Yes |
| Obese | 251,621 (29.1%) | 27,005 (3.1%) | 98,499 (11.4%) | 3,777 (0.4%) | 295,622 (34.2%) | ||||||
a: Includes: To prevent health problems, to improve my overall health, to maintain health/to stay healthy, and to prevent colds/boost immune system.
b: Includes: For healthy joints/arthritis and for bone health/build strong bones/osteoporosis.
c: DS can be used to supplement diets for people on restricted diets such as those actively participating in weight loss or those with DM
For rest of the three conditions (i.e., getting energy, losing weight and maintaining blood sugar), we found that often consumers are taking DS indiscriminately without sufficient, current and scientific knowledge on how a particular DS actually impacts human body, e.g., the use of MVMM and Vitamin B-complexes for losing weight and/or maintaining blood sugars among diabetic patients rather than their actual role of simply supplementing diet in people who are on restricted diets.
Correlation Analysis
We were interested in testing the correlations between the number of DS taken (DS Count) and age, BMI. Table 5 reports the results of Pearson correlation coefficients. Both age and BMI have significant correlations with the number of DS taken. However, the correlation between DS count and BMI are negative, which means a person having lower BMI value are likely to take more DSs. These findings are consistent with the results in the published study.22
Table 5.
Correlation analysis with the DS data with Pearson correlation coefficients
| Variable Pairs | Coefficient | P-value |
|---|---|---|
| DS Count – Age | 0.28234 | < 0.001 |
| DS Count – BMI | -0.03596 | < 0.001 |
We further used a multivariate logistic regression model to assess the impact of basic characteristics (gender, age group, race, and household income) on DS use between control and obese groups. We found that the obese group is less likely to take DS than control group. In addition, male respondents are less likely to take DS compared to female (odds ratio=0.560, CI: 0.530-0.592, p<0.001). When setting the 45-54 years old age group as the reference, younger respondents who are under 45 are less likely to take DS. On the other hand, older respondents who are over 54 are more likely to take DS. Regarding race, Mexican American, other Hispanic, non-Hispanic Black, and other race or multi-racial are all less likely to take DS compared to the reference group Non-Hispanic White. Moreover, respondents who reported the household income is $75,000 and over are more likely to take DS compared to other lower levels of household income.
Discussion
In this study, we used the NHANES data to assess the use and perceived benefits of dietary supplements among obese adults. Demographics clearly play a role in DS use. Based on the information in Table 1, female is more likely than male to use DS. With respect to age, the older the respondent was, the more likely he/she used at least one DS. The 45-54 year-old age group was the only age group in which there was no statistically significant difference in DS use between obese and control groups. Race also plays a role in DS use, with non-Hispanic Whites showing the highest percentage of DS use. Other Race or Multi-Race respondents were the only category in which there was no significance in terms of DS use. Household income is also shown to play a role in DS use. Those that have an income over $35,000 have a percentage of use of 48.08-53.47% versus those under $35,000 have a percentage of use of 32.4243.94%.
With respect to the correlation between age and BMI information and the number of DS taken, we found that both age (coefficient = 0.28, p-value < 0.001) and BMI (coefficient = -0.04, p-value <0.001) have significant correlation with the number of DS taken, indicating that older adults and those with lower BMI are likely to take more DSs. It is consistent with the results from a previous study that obese respondents reported relatively less DS use (48%) than those categorized as overweight (57%) or normal weight (56%).13 Hence, DS users, as compared to non-users, are significantly more likely to have better dietary patterns and exercises.
The overall impression from the results of this study is that obese population does not show increased usage of DS to facilitate weight loss. The results between the obese group and the control group are quite similar.
Another clear opportunity shown in the results was that most respondents took DSs on their own, as opposed to be told by the healthcare professionals. They could have learned about the DS from word of mouth, advertisement, or some sources other than a health practitioner. There is clearly an opportunity for knowledge about DSs to be passed to this population via healthcare providers. The actual benefit of each DS may be clearer if the information was from providers instead of another source.
We also investigated if the information provided in the existing knowledge base aligns with the reported use of a particular DS. We found that often consumers are taking DSs indiscriminately without sufficient, current and scientific knowledge on how a particular DS actually impacts the human body. This includes use of calcium for heart health, despite a considerable number of existing controversies (validated by various research studies) regarding the association between dietary calcium intake and risk of mortality from cardiovascular disease and its causes.23 Similarly, the use of MVMM and Vitamin B complexes for losing weight and/or maintaining blood sugar among diabetic patients rather than their actual role of simply supplementing diet in people who are on restricted diets. Similarly, use of herbs for conditions like weight loss, diabetes, heart health etc., with a false perception that they have minimal or no side effects since being natural products. Overall, we found NMCD to be reasonable in finding the relevant information for most of the ingredients/products despite a few challenges. Since NMCD has DS related information as monographs with detailed information at “ingredient” level, DS products corresponding to more than one ingredients, e.g., MVMM, Vitamin B / B-Complex need to be searched individually for each comprising ingredients for any specific information. In contrary, although Dietary Supplements Label Database (DSLD) is primarily a product level resource with plenty of multi ingredients products, the information provided is not helpful since it is (1) not specific (providing only LanguaLTM related dietary claims or uses), and (2) fragmented/distributed under various sections.
We also noted the lack of consistency of information in the DS knowledge bases. The information available to researchers/professionals is not always accurate and is not always up to date. We also acknowledge that there is no one, comprehensive DS knowledge base available to consumers. It is clear that from a consumer perspective, the DS knowledge affects the actual use of DS. In our preliminary comparison between the reported benefits for each of the major DSs and the knowledge base, there is a knowledge gap between the perceived and documented benefits of the DSs.
Limitation
A few limitations should be noted when interpreting the study results. First, the most recent data was not available at the time of the study. Second, specific usage information was not available for the entire time frame. Third, in the correlation analysis, some variables (e.g., BMI, number of DSs taken) are skewed, which may affect the correlation analysis. We attempted a few transformations (e.g., square root transformation) but have not fully resolved the data skewness problem.
Future Work
Additional studies comparing DS use in adult in the obese versus non-obese populations would provide additional information regarding DS use in the adult population as a whole and potentially provide more information regarding the use of knowledge bases. In addition, we would also like to look at medication use in the obese population using NHANES data to see what knowledge can be gained from this diverse database. We would like to also perform a study looking at specific questionnaire data related to certain health conditions and related medication use in the adult population to see what information can be gathered regarding specific diseases and how they relate to obesity. As clinical trials generate gold-standard medical evidence, we are also interested in assessing the population representativeness of DS trials and identifying the systematic biases in the eligibility criteria in the trial design using informatics methods such as Generalizability Index for Study Traits24 and its variants.18,25
Conclusions
Complementary alternative medicine, especially dietary supplements, has gained increasing popularity for many reasons including availability without prescription, price, and ease of use. Besides weight loss, there are various potential benefits for DS. Nevertheless, health consumers with limited health literacy may not adequately understand the benefits and risk of overdose for DS. In this project, we aim to gain a better understanding of the use of DS products among obese people as well as their perceived benefits of these products. We identified obese adults after combining the National Health and Nutrition Examination Survey data collected during 2003-2014. We found that there is a knowledge gap between the reported benefits of major DS by obese adults and the existing DS knowledge base and label database. This gap may inform the design of patient education material on DS usage in the future.
Table 6.
Multivariate logistic regression results
| Odds Ratio | 95% Confidence Interval | P-values | ||
|---|---|---|---|---|
| Obese or Not (reference = Control Group) | ||||
| Obese Group | 0.783 | 0.742 | 0.827 | <.0001 |
| Gender (reference = Female) | ||||
| Male | 0.560 | 0.530 | 0.592 | <.0001 |
| Age Group (reference = 45-54) | ||||
| 18-24 | 0.462 | 0.407 | 0.526 | <.0001 |
| 25-34 | 0.637 | 0.567 | 0.715 | <.0001 |
| 35-44 | 0.737 | 0.642 | 0.846 | <.0001 |
| 55-64 | 1.613 | 1.440 | 1.806 | <.0001 |
| 65-74 | 2.172 | 1.889 | 2.497 | <.0001 |
| 75 over | 2.535 | 2.219 | 2.897 | <.0001 |
| Race (reference = Non-Hispanic White) | ||||
| Mexican American | 0.510 | 0.460 | 0.565 | <.0001 |
| Other Hispanic | 0.665 | 0.593 | 0.747 | <.0001 |
| Non-Hispanic Black | 0.564 | 0.511 | 0.623 | <.0001 |
| Other Race or Multi-Racial | 0.875 | 0.769 | 0.997 | 0.0449 |
| Household Income (reference = $75,000 and over) | ||||
| $0 to $4,999 | 0.520 | 0.407 | 0.663 | <.0001 |
| $10,000 to $14,999 | 0.463 | 0.395 | 0.541 | <.0001 |
| $15,000 to $19,999 | 0.473 | 0.408 | 0.547 | <.0001 |
| $20,000 to $24,999 | 0.542 | 0.465 | 0.633 | <.0001 |
| $25,000 to $34,999 | 0.576 | 0.515 | 0.645 | <.0001 |
| $35,000 to $44,999 | 0.708 | 0.609 | 0.823 | <.0001 |
| $45,000 to $54,999 | 0.745 | 0.641 | 0.867 | 0.0002 |
| $5,000 to $9,999 | 0.452 | 0.383 | 0.533 | <.0001 |
| $55,000 to $64,999 | 0.747 | 0.641 | 0.869 | 0.0002 |
| $65,000 to $74,999 | 0.809 | 0.697 | 0.938 | 0.0054 |
| NA | 0.573 | 0.496 | 0.663 | <.0001 |
Acknowledgments
This study was partially supported by the National Center for Complementary & Integrative Health Award R01AT009457 (PI: Zhang), the National Institute on Aging award R21AG061431 (PI: He/Bian). It was also partially supported by the University of Florida Clinical and Translational Science Institute, which is supported in part by the NIH National Center for Advancing Translational Sciences under award number UL1TR001427. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
References
- 1.Pozza C, Isidori AM. Imaging in Bariatric Surgery. Springer; 2018. What’s Behind the Obesity Epidemic; pp. 1–8. [Google Scholar]
- 2.Hales CM, Fryar CD, Carroll MD, Freedman DS, Ogden CL. Trends in Obesity and Severe Obesity Prevalence in US Youth and Adults by Sex and Age, 2007-2008 to 2015-2016. JAMA. 2018;319(16):1723–5. doi: 10.1001/jama.2018.3060. PMC5876828. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Church T, Martin CK. The Obesity Epidemic: A Consequence of Reduced Energy Expenditure and the Uncoupling of Energy Intake? Obesity (Silver Spring) 2018;26(1):14–6. doi: 10.1002/oby.22072. [DOI] [PubMed] [Google Scholar]
- 4.Kuntz B, Lampert T. Socioeconomic factors and obesity. Dtsch Arztebl Int. 2010;107(30):517–22. doi: 10.3238/arztebl.2010.0517. PMC2925342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.CDC. Adult Obesity Facts. [October 31, 2018]. Available from: https://www.cdc.gov/obesity/data/adult.html. [Google Scholar]
- 6.Blackburn GL, Walker WA. Science-based solutions to obesity: what are the roles of academia, government, industry, and health care? Am J Clin Nutr. 2005;82(1 Suppl):207S-10S. doi: 10.1093/ajcn/82.1.207S. [DOI] [PubMed] [Google Scholar]
- 7.Schelbert KB. Comorbidities of obesity. Primary Care: Clinics in Office Practice. 2009;36(2):271–85. doi: 10.1016/j.pop.2009.01.009. [DOI] [PubMed] [Google Scholar]
- 8.Scott KM, Bruffaerts R, Simon GE, Alonso J, Angermeyer M, de Girolamo G, Demyttenaere K, Gasquet I, Haro JM, Karam E, Kessler RC, Levinson D, Medina Mora ME, Oakley Browne MA, Ormel J, Villa JP, Uda H, Von Korff M. Obesity and mental disorders in the general population: results from the world mental health surveys. Int J Obes (Lond) 2008;32(1):192–200. doi: 10.1038/sj.ijo.0803701. PMC2736857. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fujioka K. Current and emerging medications for overweight or obesity in people with comorbidities. Diabetes Obes Metab. 2015;17(11):1021–32. doi: 10.1111/dom.12502. PMC4744746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Biener A, Cawley J, Meyerhoefer C. The High and Rising Costs of Obesity to the US Health Care System. J Gen Intern Med. 2017;32(Suppl 1):6–8. doi: 10.1007/s11606-016-3968-8. PMC5359159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Burgess E, Hassmen P, Pumpa KL. Determinants of adherence to lifestyle intervention in adults with obesity: a systematic review. Clin Obes. 2017;7(3):123–35. doi: 10.1111/cob.12183. [DOI] [PubMed] [Google Scholar]
- 12.Saper RB, Eisenberg DM, Phillips RS. Common dietary supplements for weight loss. Am Fam Physician. 2004;70(9):1731–8. [PubMed] [Google Scholar]
- 13.Bailey RL, Gahche JJ, Lentino CV, Dwyer JT, Engel JS, Thomas PR, Betz JM, Sempos CT, Picciano MF. Dietary supplement use in the United States, 2003-2006. J Nutr. 2011;141(2):261–6. doi: 10.3945/jn.110.133025. PMC3021445. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Troxler DS, Michaud PA, Graz B, Rodondi PY. Exploratory survey about dietary supplement use: a hazardous and erratic way to improve one’s health? Swiss Med Wkly. 2013;143:w13807. doi: 10.4414/smw.2013.13807. [DOI] [PubMed] [Google Scholar]
- 15.CDC. National Health and Nutrition Examination Survey Data: U S Department of Health and Human Services, Centers for Disease Control and Prevention. [October 31, 2018]. Available from: http://www.cdc.gov/nchs/nhanes.htm. [Google Scholar]
- 16.Vladutiu CJ, Ahrens KA, Verbiest S, Menard MK, Stuebe AM. Cardiovascular Health of Mothers in the United States: National Health and Nutrition Examination Survey 2007-2014. J Womens Health (Larchmt) 2018 doi: 10.1089/jwh.2018.7204. [DOI] [PubMed] [Google Scholar]
- 17.He Z, Wang S, Borhanian E, Weng C. Assessing the Collective Population Representativeness of Related Type 2 Diabetes Trials by Combining Public Data from ClinicalTrials.gov and NHANES. Stud Health Technol Inform. 2015;216:569–73. PMC4586087. [PMC free article] [PubMed] [Google Scholar]
- 18.He Z, Ryan P, Hoxha J, Wang S, Carini S, Sim I, Weng C. Multivariate analysis of the population representativeness of related clinical studies. J Biomed Inform. 2016;60:66–76. doi: 10.1016/j.jbi.2016.01.007. PMC4837055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.He Z, Charness N, Bian J, Hogan WR. Assessing the Comorbidity Gap between Clinical Studies and Prevalence in Elderly Patient Populations. IEEE EMBS Int Conf Biomed Health Inform. 2016;2016:136–9. doi: 10.1109/BHI.2016.7455853. PMC5058342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Rizvi RF, Adam TJ, Lindemann EA, Vasilakes J, Pakhomov SV, Bishop JR, Melton GB, Zhang R. Comparing Existing Resources to Represent Dietary Supplements. AMIA Jt Summits Transl Sci Proc. 2018;2017:207–16. PMC5961776. [PMC free article] [PubMed] [Google Scholar]
- 21.Natural Medicine Comprehensive Database (NMCD) [December 2, 2018]. Available from: https://naturalmedicines.therapeuticresearch.com. [Google Scholar]
- 22.Cowan AE, Jun S, Gahche JJ, Tooze JA, Dwyer JT, Eicher-Miller HA, Bhadra A, Guenther PM, Potischman N, Dodd KW, Bailey RL. Dietary Supplement Use Differs by Socioeconomic and Health-Related Characteristics among U.S. Adults, NHANES 2011(-)2014. Nutrients. 2018;10(8) doi: 10.3390/nu10081114. PMC6116059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wang X, Chen H, Ouyang Y, Liu J, Zhao G, Bao W, Yan M. Dietary calcium intake and mortality risk from cardiovascular disease and all causes: a meta-analysis of prospective cohort studies. BMC Med. 2014;12:158. doi: 10.1186/s12916-014-0158-6. PMC4199062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Weng C, Li Y, Ryan P, Zhang Y, Liu F, Gao J, Bigger JT, Hripcsak G. A distribution-based method for assessing the differences between clinical trial target populations and patient populations in electronic health records. Appl Clin Inform. 2014;5(2):463–79. doi: 10.4338/ACI-2013-12-RA-0105. PMC4081748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sen A, Chakrabarti S, Goldstein A, Wang S, Ryan PB, Weng C. GIST 2.0: A scalable multi-trait metric for quantifying population representativeness of individual clinical studies. J Biomed Inform. 2016;63:325–36. doi: 10.1016/j.jbi.2016.09.003. PMC5077682. [DOI] [PMC free article] [PubMed] [Google Scholar]