Spam on health and pharmaceutical topics are received. Weight Problems and Spam for Weight Loss Products.

Transcription

1 Original Article Weight Problems and Spam for Weight Loss Products Joshua Fogel, PhD, and Sam Shlivko, BS Background: This study focuses on young adult behaviors with regard to spam s that sell weight loss products. Methods: Participants (N 200) with and without weight were asked if they received, opened, and bought products from spam about weight loss topics in the past year. Psychological factors of self-esteem and perceived stress were measured. Results: Those with weight had significantly greater percentages than those without weight for receiving (87.7% vs. 73.3%, P 0.02), opening (41.5% vs. 17.8%, P 0.001), and buying products (18.5% vs. 5.2%, P 0.003). In the multivariate logistic regression analyses, weight were significantly associated with receiving (: 3.39, 95% CI: 1.31, 8.82), opening (: 3.10, 95% CI: 1.53, 6.29), and buying products (: 3.38, 95% CI: 1.16, 9.82). Conclusion: Physicians should consider discussing with patients the potential risks of opening and/or purchasing weight loss products from spam s. Key Words: body weight, electronic mail, internet, obesity, weight loss Spam on health and pharmaceutical topics are received by over 80% of individuals. 1 In a 2007 survey, the Pew Internet and American Life Project reported that most users are bothered by spam with 18% believing that it is a big problem and 51% believing that it is annoying. 2 Another study reported that 81% of individuals strongly dislike spam and 14% simply just dislike spam. 3 Spam s are sent with the intent for recipients to open them and to be interested in the message content of that spam . Purchases from spam solicitations range from as low as 4% to as high as 66% depending upon the study. 1 5 Older From the Department of Economics, Brooklyn College, Brooklyn; and New York Law School, New York, NY. Reprint requests to Joshua Fogel, PhD, Department of Economics, Brooklyn College of the City University of New York, 218A, 2900 Bedford Avenue, Brooklyn, NY [email protected] Accepted April 14, Copyright 2010 by The Southern Medical Association /0 2000/ retirement-age individuals are more likely to purchase from solicitations in spam than college-age young adults. 3 Almost one-third of spam consists of health-related messages. 6 Spam marketing of health products appeals to the perceived health needs of individuals. Weight loss topics are of great importance to young adults and spam s about weight loss are commonly sent to young adults. There are also psychological concerns related to dissatisfaction with one s weight that can include low self-esteem 7 and increased perceived stress. 8,9 These psychological concerns may make young adults susceptible to buy products from spam weight loss solicitations. To our knowledge, we are the first to study young adult responses to spam selling weight loss products. We have four hypotheses. First, the presence of weight among young adults is associated with receiving spam about weight loss topics at greater levels than those without weight. Spam is not always randomly sent to all individuals with an address. Individuals with weight may have previously provided their address when they requested information or purchased weight loss products from internet vendors. These addresses may be used by this particular company or sold to other companies who now send spam to these individuals. Individuals who do not have weight are not as likely to have provided their address to internet vendors selling weight loss products. Second, the presence of weight among young adults is associated with opening spam about weight loss topics at greater levels than those without weight due to greater interest in weight loss. Third, the presence of weight among young adults is associated with buying products from spam about weight loss topics at greater levels than those without weight due to greater interest in weight loss. Fourth, the psychological concerns of low self-esteem and increased perceived stress, which are often associated with being overweight, is associated with the receiving of, opening of, and buying products from spam about weight loss topics. Materials and Methods Participants and Procedures Participants were 200 students from a 4-year undergraduate commuter college in New York City. Of 212 students Southern Medical Journal Volume 103, Number 1, January

2 Fogel and Shlivko Spam and Weight Loss approached, 200 completed surveys for a response rate of 94.3%. The sample was a convenience sample. Participants were approached to complete an anonymous survey in classrooms, the school cafeteria, the library, and other public places. Informed consent was obtained. The survey was exempt from Institutional Review Board review and was conducted in accordance with the ethical principles of the Declaration of Helsinki. Data collection occurred during May Measures Demographics. Demographic variables included age (years), sex, race/ethnicity (white, non-white), hours of internet use (daily), and number of spam s received (daily). Weight item. Participants were asked, Do you believe that you have weight? with choices of yes or no. Spam items. These items were: 1) Did you receive spam about weight loss in the past year? 2) If yes, did you open and read the ? 3) If you opened and read the , did you buy anything from the website provided? Psychological Questionnaires Rosenberg Self-Esteem Scale. The Rosenberg Self-Esteem Scale 10 is a reliable and valid measure. It includes 10 items measured on a Likert-style scale ranging from 1 strongly disagree to 4 strongly agree. Five items are reverse coded. Higher scores indicate greater self-esteem. Cronbach alpha reliability ranges from 0.77 to ,11 In this sample, Cronbach alpha reliability was Perceived Stress Scale. The Perceived Stress Scale 12 is a reliable and valid measure. It includes 10 items measured on a Likert-style scale ranging from 0 never to 4 very often. Four items are reverse coded. Higher scores indicate greater perceived stress. Cronbach alpha reliability ranges from 0.80 to In this sample, Cronbach alpha reliability was Table 1. Demographic characteristics of a sample of 200 individuals regarding their weight status a Variable No weight M (SD) or % (no.) (n 135) Yes weight M (SD) or % (no.) (n 65) P Age (yr) b 20.9 (2.04) 21.0 (1.91) 0.69 Sex c Men 40.7% (55) 24.6% (16) 0.03 Women 59.3% (80) 75.4% (49) Race/ethnicity c 0.60 White 42.2% (57) 46.2% (30) Non-white 57.8% (78) 53.8% (35) Hours internet use 3.8 (2.39) 4.2 (2.57) 0.13 (daily) b No. spam s (daily) d 29.7 (71.24) 25.1 (35.74) 0.39 P-values for hours of internet use are from logarithmically transformed variables due to presence of skewness. Nonlogarithmically transformed values are shown for the mean and standard deviation values for ease of understanding. Number of spam s received, although skewed, was not logarithmically transformed due to eight individuals with values of 0, precluding a logarithmic transformation. a M, mean; SD, standard deviation. b Analysis of variance performed. c Pearson 2 test performed. d Mann-Whitney test performed. Statistical Analyses When appropriate, analysis of variance (ANOVA), the Pearson chi-square test, or the Mann-Whitney test were used to compare those with and without weight. Separate ANOVA and analysis of covariance (ANCOVA) were conducted with the psychological dependent variables of the Rosenberg Self-Esteem Scale and the Perceived Stress Scale. Covariates included age, sex, race/ethnicity, hours of internet use (daily), and number of spam s received (daily). Pearson chi-square analyses were used to compare the responses of those with and without weight to the separate questions of receiving, opening, and purchasing from spam about weight loss topics. Lastly, we conducted a series of logistic regression analyses with three different outcome variables for receiving, opening, and purchasing from spam about weight loss topics. For each outcome variable, three models were conducted. The first model was conducted with the independent variable alone. The second model was conducted with the independent variable and the relevant demographic variables. The third model was conducted with the independent variable, the relevant demographic variables, and the psychological variables of self-esteem and perceived stress. SPSS version (SPSS, Inc., Chicago, IL) was used for all analyses, except for the calculation of Cohen s d, which was found using an effect size calculator. 14 Results Table 1 shows comparisons for those with (32.5%; n 65) and without (67.5%; n 135) weight as pertaining to the demographic characteristics of the sample. The average age was 21 years for both groups. There were a significantly greater percentage of women than men with weight. Slightly more than half the sample was non-white. Each of the groups averaged four hours of daily internet use. Although those without weight reported slightly more spam received daily (30 versus 25), this did not significantly differ between the groups. Table 2 shows comparisons for weight and spam behaviors. Those with weight were significantly more likely than those without weight to have received spam about weight loss topics, opened spam about weight loss topics, and bought from spam Southern Medical Association

3 Original Article Table 2. Weight and spam behaviors Variable No weight % (no.) (n 135) Yes weight % (no.) (n 65) P Received weight loss No 26.7 (36) 12.3 (8) 0.02 Yes 73.3 (99) 87.7 (57) Opened weight loss No 82.2 (111) 58.5 (38) Yes 17.8 (24) 41.5 (27) Bought from weight loss No 94.8 (128) 81.5 (53) Yes 5.2 (7) 18.5 (12) about weight loss topics. Also, 18.5% [men: 3/16, 18.8%; women: 9/49, 18.4%] of those with weight bought from spam about weight loss topics while only about 5% of those without weight bought from spam about weight loss topics. With regard to comparisons for the psychological variables, those with weight had lower self-esteem levels (M 29.2, SD 6.33) than those without weight (M 31.4, SD 4.88). This was significantly different with ANOVA (P 0.01) and also with ANCOVA (P 0.02) when adjusting for the covariates. There was a small-medium effect size (d 0.38). Also, those with weight had greater perceived stress levels (M 21.1, SD 6.38) than those without weight (M 18.3, SD 5.82). This was significantly different with ANOVA (P 0.002) and also with ANCOVA (P 0.005) when adjusting for the covariates. There was a medium effect size (d 0.46). Table 3 shows the logistic regression analyses for receiving spam about weight loss topics. In Model 1 those with weight had a significant odds ratio of 2.59 as compared to those without weight for receiving spam about weight loss topics. In Model 2 that included demographics, the odds ratio for receiving spam e- mail for those with weight increased to Also, the number of spam s received was significantly associated with an increase in receiving spam about weight loss topics. In Model 3 that included the psychological variables, both the presence of weight and number of spam s had similar odds ratios as in Model 2 and were still significantly associated with an increase in receiving weight loss spam s. Also, increasing age was associated with receiving less spam s about weight loss topics. None of the psychological variables were associated with receiving weight loss spam s. The Figure includes the odds ratios for presence of weight for Model 3. Table 3. Logistic regression analyses for receiving weight loss spam a Variables model 1 model 2 model 3 Weight No Yes 2.59 ( ) b 3.24 ( ) b 3.39 ( ) b Age 0.83 ( ) b 0.81 ( ) b Sex Men Women 1.11 ( ) 1.18 ( ) Race White Non-white 0.95 ( ) 0.98 ( ) Hours internet 0.91 ( ) 0.91 ( ) No. spam s 1.04 ( ) c 1.04 ( ) c Rosenberg 0.98 ( ) self-esteem Perceived stress 0.96 ( ) a, odds ratio; CI, confidence interval. b P c P Table 4 shows the logistic regression analyses for opening the spam s. In all 3 models, only the presence of weight was associated with significantly increased odds for opening the s. The odds ratios ranged from 3.10 to 3.29 depending upon the model (also see Fig.). Table 5 shows the logistic regression analyses for buying from spam about weight loss topics. In all 3 models, the presence of weight was associated with significantly increased odds for buying products advertised in the spam s. For Models 1 and 2, the odds ratios were more than 4 for doing so as compared to those without weight. In Model 3, the presence of weight had an odds ratio of more than 3 for buying the advertised weight loss products. Also, increased perceived stress was associated with buying habits (Table 3) (also see Fig.). Discussion We found strong support for three of our four hypotheses. Hypothesis 1 was supported where we found that the presence of weight among young adults is associated with increased odds for receiving spam about weight loss topics. Hypothesis 2 was supported where we found that the presence of weight among young adults is associated with increased odds for opening the weight loss spam s. Hypothesis 3 was supported where we found that the presence of weight among young adults is associated with increased odds for buying weight Southern Medical Journal Volume 103, Number 1, January

4 Fogel and Shlivko Spam and Weight Loss Table 4. Logistic regression analyses for opening weight loss spam a Table 5. Logistic regression analyses for buying products from weight loss spam a Variables model 1 model 2 model 3 Variables model 1 model 2 model 3 Weight No Yes 3.29 ( ) b 3.29 ( ) c 3.10 ( ) d Age 1.10 ( ) 1.12 ( ) Sex Men Women 1.74 ( ) 1.66 ( ) Race White Non-white 1.35 ( ) 1.29 ( ) Hours Internet 0.70 ( ) 0.91 ( ) No. spam s 1.00 ( ) 1.00 ( ) Rosenberg 1.01 ( ) self-esteem Perceived stress 1.03 ( ) a, odds ratio; CI, confidence interval. b P c P Weight No Yes 4.14 ( ) b 4.23 ( ) b 3.38 ( ) c Age 1.11 ( ) 1.23 ( ) Sex Men Women 1.05 ( ) 0.82 ( ) Race White Non-white 1.47 ( ) 1.15 ( ) Hours internet 0.83 ( ) 0.95 ( ) No. spam s 1.00 ( ) 1.00 ( ) Rosenberg 1.03 ( ) self-esteem Perceived stress 1.14 ( ) c a, odds ratio; CI, confidence interval. b P c P d P loss products from spam . However, we only found partial support for hypothesis 4 where increased perceived stress is associated with buying products. However, perceived stress was not associated with receiving or opening the spam s. Also, self-esteem was not related to receiving, opening, or buying the advertised products. Prevalence of Spam on Weight Loss Topics Spam on weight loss topics is quite common. Although significantly differing between those with and without weight, spam s about weight loss arrived in more than 70% of the inboxes of all participants. The opening of this spam was still somewhat of interest, where almost 18% of those without weight and more than 40% of those with weight opened the spam . With regard to buying advertised products, there Odds Ratio Fig Receiving Opening Buying Spam about weight loss topics. No Weight Problems Yes Weight Problems again was much more interest among those with weight (18.5% purchased products). There were also similar percentages for buying among men and women. However, only about 5% from those without weight bought products. It is interesting to note that, although a very small percentage, some individuals without weight bought products from the spam . A possible reason is that, although they do not have a problem with their weight, these individuals still believe that it would be worthwhile to lose some weight. Also, our overall buying behavior from spam from both the weight and no weight groups was 9.5% (19/200), which is slightly lower than the 13% reported from a six country survey for purchasing any health or pharmaceutical products from spam . 1 The National Physical Activity and Weight Loss Survey reports the prevalence of nonprescription weight-loss supplement use within the past 12 months. 15 Among its young adult group aged years, the prevalence for women was 16.7% and 10.1% for men. Our buying percentage of 18.4% for women is very similar. However, our buying percentage of 18.8% for men is slightly higher. Psychological Variables and Weight Concerns As was expected for the psychological variables, we found lower self-esteem levels and increased perceived stress levels among those with weight as compared to those without weight. These psychological variables are not just restricted to these particular psychological Southern Medical Association

5 Original Article constructs. They can relate to other psychological and physical health concerns too. For example, among college students with weight, lower self-esteem was significantly related to a number of measures of body image dissatisfaction. 7 Also, increased stress negatively affected those who were trying to restrain their food intake and who, due to their stress, ate more food. 16 Spam and Behavior Those with weight had similar odds ratios of greater than 3 for the models that also included the demographic and psychological variables and the 3 separate outcomes of receiving, opening, and buying from the spam e- mails (Fig.). A literature search did not yield any studies on spam use and weight loss. However, there are a number of studies which show that for those seeking treatment for weight loss, counseling done through is effective in helping individuals lose weight. We suggest that spam may appeal to those with weight as a potential venue for helping them address their weight. This pattern of buying behavior from spam is of concern, as these individuals are apparently not seeking or are not satisfied with the evidence-based treatments available from physicians, psychologists, dieticians, nurses, exercise physiologists, or other health care providers. Another area of concern is with regard to the content of the weight loss products that are bought by these consumers from this type of spam . A review of products sold by online pharmacies indicates that pharmaceutical and herbal products sold over the internet can range in quality from poor to excellent. 21 One concern is that spam is potentially not any better than content sold by online pharmacies and that the quality too can range from harmless to potentially dangerous nonprescription products. A second concern is that some spam vendors advertise products that would require a prescription in the United States and are sold to consumers without requiring a prescription. Our results among young adults show that an increased number of spam s received is associated with receiving spam about weight loss topics, while the number of hours of internet use is not associated with receiving spam about weight loss topics. This is similar to the significant association between the number of spam s received and the health spam received and to the lack of a relationship for hours of internet use and health spam received among college-age and working-age individuals. 3 In the analyses for opening spam about weight loss topics, no demographic or psychological variables were associated with opening these s. However, increased perceived stress was associated with increased odds for buying from the spam s. Apparently, just as consumers shop for consumer goods as a form of stress relief, 22 stress relief may be an additional reason for purchasing weight loss products from spam. Physician Implications The American Medical Association strongly encourages physicians to assess weight and to discuss with patients the health implications of being overweight or obese. 23 Patients will not typically initiate a discussion with their physician about weight concerns. Even when patients are already purchasing nonprescription weight-loss supplements, less than one-third of young adults discuss these supplements with a physician or other health care professional. 15 It is also not so simple to effectively counsel patients to lose weight. One approach found effective by many primary care physicians is to focus on improving a patient s general health habits and wellness instead of focusing on a particular weight. 24 Along with assessing and possibly counseling and/or treating patients for weight topics, it would be worthwhile to inquire about weight loss supplement use and also the source of the supplements. Study Limitations and Future Research The study has some potential limitations. First, these data were collected at one university and may not be representative of a national sample. Second, it may have been informative to obtain the body mass index of the participants to determine if being overweight was related to this buying behavior. However, weight perception is often quite subjective and many individuals who have appropriate weight will still seek to lose just a few pounds, so objective overweight status is not critical to understanding this spam buying behavior. Third, as this is the first such study, future research would be useful to conduct similar research to replicate these study findings with a larger sample size in a national sample. Fourth, it would be useful in future research to understand the actual products purchased from spam (ie herbal products, prescription medication, etc) to determine the actual potential risk from using such products. Fifth, future studies should also assess whether there are adverse events from the weight loss products purchased online. Conclusion Spam s about weight loss are being looked at and purchased from, especially among those who are sensitive to weight concerns. Clearly, those who are sending these spam marketing s have a receptive audience. Physicians, psychologists, dieticians, nurses, exercise physiologists, or other health care providers who assess, counsel, and treat individuals for weight concerns should discuss with their patients the potential risks of opening and/or purchasing from spam e- mail. They should emphasize to their patients the importance of working together with a health care professional in coordinating care when considering the use of weight loss products. It may even be useful for physicians to ask all their patients if they are purchasing health products from spam . Future research should study if health care profes- Southern Medical Journal Volume 103, Number 1, January

6 Fogel and Shlivko Spam and Weight Loss sionals discussions with patients who purchase from spam results in changed purchasing behavior from spam among these patients. References 1. Forrester Research, Business Source Alliance. Consumer Attitudes Toward Spam in Six Countries Available at: country/research%20and%20statistics/ /media/87f4037f2b70 44ECBF448A227B84BE86.ashx. Accessed January 21, Fallows D. Spam Available at: PIP_Spam_May_2007.pdf. Accessed July 12, Grimes GA, Hough MG, Signorella ML. end users and spam: relations of gender and age group to attitudes and actions. Comput Human Behav 2007;23: Mailshell. Results of the SpamCatcher Attitude Survey Available at: Accessed June 02, Marshal. Sex, Drugs and Software Lead Spam Purchase Growth Available at: 748&thesection news. Accessed January 21, Gernburd P, Jadad AR. Will spam overwhelm our defenses? Evaluating offerings for drugs and natural health products. PLoS Med 2007;4:e Lowery SE, Robinson Kurpius SE, Befort C, et al. Body image, selfesteem, and health-related behaviors among male and female first year college students. J Coll Stud Dev 2005;46: Metz U, Welke J, Esch T, et al. Perception of stress and quality of life in overweight and obese people implications for preventive consultancies in primary care. Med Sci Monit 2009;15:PH1 PH6. 9. Nishitani N, Sakakibara H. Relationship of obesity to job stress and eating behavior in male Japanese workers. Int J Obes 2006;30: Rosenberg M. Conceiving the Self. Malabar, FL, Krieger, Blascovich J, Tomaka J. Measures of Self-Esteem, in Robinson JP, Shaver PR, Wrightsman LS (eds): Measures of Personality and Social Psychological Attitudes. Ann Arbor, Institute for Social Research, 1993, ed 3, pp Cohen S, Williamson G. Perceived Stress in a Probability Sample of the United States, in Spacapan S, Oskamp S (eds): The Social Psychology of Health: Claremont Symposium on Applied Social Psychology. Newbury Park, CA, Sage, 1988, pp SPSS Version 15. Chicago, SPSS, Becker. Effect Size Calculators Available at: lbecker/psy590/escalc3.htm. Accessed August 9, Blanck HM, Serdula MK, Gillespie C, et al. Use of nonprescription dietary supplements for weight loss is common among Americans. JAm Diet Assoc 2007;107: Wardle J, Steptoe A, Oliver G, et al. Stress, dietary restraint and food intake. J Psychosom Res 2000;48: Harvey-Berino J, Pintauro S, Buzzell P, et al. Effect of internet support on the long-term maintenance of weight loss. Obes Res 2004;12: Tate DF, Jackvony EH, Wing RR. Effects of Internet behavioral counseling on weight loss in adults at risk for type 2 diabetes: a randomized trial. JAMA 2003;289: Tate DF, Jackvony EH, Wing RR. A randomized trial comparing human counseling, computer-automated tailored counseling, and no counseling in an Internet weight loss program. Arch Intern Med 2006; 166: Tate DF, Wing RR, Winett RA. Using internet technology to deliver a behavioral weight loss program. JAMA 2001;285: Fogel J. Consumers and Purchases of Health Products Over the Internet, in Saito F (ed): Consumer Behavior. Hauppauge, NY, Nova Science, 2009, pp Hama Y. Shopping as a coping behavior for stress. Jpn Psychol Res 2001;43: Lyznicki JM, Young DC, Riggs JA, et al. Obesity: assessment and management in primary care. Am Fam Physician 2001;63: Sussman AL, Williams RL, Leverence R, et al. The art and complexity of primary care clinicians? Preventive counseling decisions: obesity as a case study. Ann Fam Med 2006;4: Southern Medical Association