Journal of Health Communication, 20:105–111, 2015 Copyright # Taylor & Francis Group, LLC ISSN: 1081-0730 print/1087-0415 online DOI: 10.1080/10810730.2014.914607
The Association Between Fruit and Vegetable Intake, Knowledge of the Recommendations, and Health Information Seeking Within Adults in the U.S. Mainland and in Puerto Rico ´ N COLO ´ N-RAMOS1, LILA J. FINNEY RUTTEN2, RICHARD P. MOSER3, VIVIAN COLO ´ N-LOPEZ4, URIYOA 5 6 ANA P. ORTIZ , and AMY LAZARUS YAROCH 1
Department of Global Health, School of Public Health and Health Services, George Washington University, Washington, District of Columbia, USA 2 Department of Health Sciences Research, Division of Epidemiology, and Population Health Science Program, Center for the Science of Health Care Delivery, Mayo Clinic, Rochester, Minnesota, USA 3 Behavioral Research Program, National Cancer Institute, Bethesda, Maryland, USA 4 Department of Health Services Administration, Graduate School of Public Health, University of Puerto Rico, San Juan, Puerto Rico 5 Department of Biostatistics and Epidemiology, Graduate School of Public Health, Medical Sciences Campus, University of Puerto Rico, San Juan, Puerto Rico 6 Gretchen Swanson Center for Nutrition, Omaha, Nebraska, USA
Health information correlates of fruit and vegetable intake and of knowledge of the fruit and vegetable recommendations were examined using bivariate and multivariate regressions with data from the 2007–2008 U.S. National Cancer Institute’s Health Information National Trends Survey in the United States and in Puerto Rico. Residents from Puerto Rico had the lowest reported fruit and vegetable intake and the lowest knowledge of the recommended servings of fruits and vegetables to maintain good health, compared with U.S. Hispanics, non-Hispanic Whites, and Blacks. Sixty-seven percent of Puerto Rican residents and 62% of U.S. Hispanics reported never seeking information on health or medical topics. In multivariate analysis, those who never sought information on health or medical topics reported significantly lower fruit and vegetable intake (coefficient ¼ 0.24; 95% CI [0.38, 0.09]), and were less likely to know the fruit and vegetable recommendations (OR ¼ 0.32; 95% CI [0.20, 0.52]), compared with those who obtained information from their health care providers. Health promotion initiatives in the United States and Puerto Rico have invested in mass media campaigns to increase consumption of and knowledge about fruit and vegetables, but populations with the lowest intake are less likely to seek information. Strategies must be multipronged to address institutional, economic, and behavioral constraints of populations who do not seek out health information from any sources.
The 2010 Dietary Guidelines urges American adults to increase fruit and vegetable (FV) intake (U.S. Department of Agriculture, 2010). Diets high in these nutrient-rich, low-calorie foods have been associated with better weight management (Rolls, Ello-Martin, & Tohill, 2004) and decreased risk of chronic diseases (U.S. Department of Agriculture, 2010). Despite educational campaigns to tout knowledge of these recommendations (Produce for Better Health & Centers for Disease Control and Prevention, n.d.), most Americans consume less than three servings a day (Thompson et al., 2005). The low consumption is possibly Address correspondence to Uriyoa´n Colo´n-Ramos, Department of Global Health, Milken Institute School of Public Health, George Washington University, 950 New Hampshire Avenue #410, Washington, DC 20052, USA. E-mail: uriyoan@ gwu.edu
the result of economic and institutional constraints, as well as behavioral and sociocultural barriers (Kraak, Story, & Swinburn, 2013; World Health Organization, 2005). As part of the latter, knowledge of the FV recommendation can be an important component of multipronged interventions to improve FV intake, because it has been associated with raised awareness and behavioral intent (National Fruit & Vegetable Alliance, 2010), and with FV intake among adults (Colon-Lopez et al., 2013; Shaikh, Yaroch, Nebeling, Yeh, & Resnicow, 2008). However, little is known about how populations with low FV consumption attain knowledge of the FV recommendation. In the U.S. mainland, there are evident dietary disparities, with U.S. Hispanics reporting higher consumption compared with U.S. non-Hispanic Whites and Blacks (Blanck, Gillespie, Kimmons, Seymour, & Serdula, 2008; ColonRamos et al., 2009; Thompson et al., 2005). In Puerto Rico,
U. Colo´n-Ramos et al.
106 a commonwealth of the United States where 99% of the residents are Hispanic (U.S. Census Bureau, 2010), only 13.7% of the population consumes at least five servings of FV a day, compared with 24.3% of the general U.S. population (Centers for Disease Control and Prevention, 2010). Understanding health information correlates of these dietary disparities and of knowledge of the FV recommendation among racial=ethnic groups in the U.S. mainland and in Puerto Rico will shed light on how to best promote FV intake in these diverse populations. We used data collected via the U.S. National Cancer Institute’s Health Information National Trends Survey in Puerto Rico (HINTS-Puerto Rico) and U.S. mainland (HINTS 2007–2008) (Tortolero-Luna et al., 2010) to elucidate potential health information sources related to FV consumption in Puerto Rico and in the U.S. mainland (non-Hispanic Whites, Blacks, and mainland Hispanics).
Method HINTS is a national survey designed to track population trends in cancer-relevant behavior, knowledge, and attitudes and assess the American public’s use of and experiences with health information (National Cancer Institute, n.d.). Data were collected from HINTS 2007–2008 in the U.S. mainland, and a special data collection effort that used the HINTS 2007–2008 items in Puerto Rico (HINTS-Puerto Rico) in 2009 (Tortolero-Luna et al., 2010). The HINTS 2007–2008 instrument includes previous iterations of HINTS questions (about 50% of the items) and new questions that were developed and reviewed by working groups at the National Cancer Institute. The instrument underwent three rounds of cognitive tests. The instrument was also translated into Spanish by a native Spanish speaker with extensive experience in survey translation. The instrument was reviewed, edited, and pilot-tested for the island population by the HINTS Puerto Rico team. Data Collection and Response Rates HINTS 2007–2008 employed a mixed mode, dual-frame design (n ¼ 7,674) including list-assisted, random-digit dial computer-assisted telephone interview (n ¼ 4,092) and a self-administered mailed survey that used a postal frame of U.S. addresses (n ¼ 3,582). Using the same instrument (translated to Spanish), HINTS-Puerto Rico was administered solely as a random digit dialing survey (n ¼ 639); therefore, to keep mode consistent, our analyses were restricted to the random digit dialing sample for the U.S. mainland. In the U.S. mainland, the random digit dialing survey was conducted from January to April 2008. The random digit dialing household screener response rate was 42.4%, and the response rate for extended interview was 57.2%, with an overall response rate of 24.2%. In Puerto Rico, a stratified sampling frame was used representing the eight geographic regions of the Puerto Rico Department of Health. The survey in Puerto Rico was conducted by bilingual interviewers (Spanish and English) between April and June 2009, with a response rate of 76.3% and a total of 639 interviews
completed. Greater details on the HINTS 2007–2008 sample and sampling design, and HINTS-Puerto Rico have been published elsewhere (Cantor et al., 2009; Tortolero-Luna et al., 2010). Study Variables Sociodemographic characteristics—sex, education, age, and household income—were collected. Diet-related variables and health information variables were assessed. Diet-Related Variables Two questions were used to asses FV intake: 1. ‘‘How many servings of fruits do you usually eat or drink each day? Think of a serving as being about 1 medium piece, or 1=2 cup of fruit, or 3=4 of cup of fruit juice.’’ 2. ‘‘How many servings of vegetables do you usually eat or drink each day? Think of a serving as being about 1 cup of raw leafy vegetables, 1=2 cup of other cooked or raw vegetables, or 3=4 cup of vegetable juice.’’ Both questions were open-ended, and participants were asked to write in a number for servings each day. Validity for these questions was established for the U.S. mainland population (Yaroch et al., 2012). Knowledge of FV recommendation was assessed by asking, ‘‘How many servings of fruits and vegetables do you think the average adult should eat each day for good health?’’ Responses were open-ended and calculated as dichotomous (those who responded anywhere between 5–13 servings=day vs. those who responded outside of that range) and continuous. Health Information Variables Participants were asked the following: 1. ‘‘Have you ever looked for information about health or medical topics from any source?’’ (Y=N) 2. ‘‘The most recent time you looked for information about health or medical topics, where did you go first?’’ (Twelve choices were provided with the option to write in ‘other’ choices) Participant responses were then grouped into the following: Internet, doctor or health care provider, printed materials (library, magazines, newspapers, books, brochures, pamphlets), and other (cancer organization, family, friend or coworker, telephone information number, or alternative practitioner). Statistical Analyses We used SAS-callable SUDAAN version 9.0 to account for the complex survey design of HINTS 2007–2008 and HINTS-Puerto Rico. All analyses used sample weights to obtain population-level estimates of the populations in the mainland U.S. and Puerto Rico, respectively. Weighted frequencies were calculated for Puerto Rico and by race= ethnicity in the United States (U.S. non-Hispanic Whites, U.S. non-Hispanic Blacks, and U.S. Hispanics). The majority of respondents in HINTS-Puerto Rico (99%) were of
Health Information Correlates of Diet in Puerto Rico and United States Hispanic=Latino origin. Also, almost half of the respondents in HINTS-Puerto Rico (44.6%) refused to report income or had missing values, so we report them as refused=missing. We first summarized weighted frequencies for each of the variables stratified by race=ethnicity, and calculated a p value derived from a chi-square statistic (for categorical variables) or Wald F statistic (for continuous values) to assess differences across distribution of sociodemographic and health information characteristics by racial=ethnic groups. These tests were also used to identify potential associations for each variable of interest with FV intake or knowledge of FV recommendation (the latter results are not shown). Median and means of FV intake and of responses to the number of servings needed for good health were examined. Kurtosis and skewness values were examined to assess assumption of normal distribution for both variables. We also dichotomized the responses for the knowledge variable (knowledge on the number of servings needed for good health) into two categories: 5–13 servings=day and anything outside of that range. We used the more comprehensive range of 5–13 servings=day to allow for inclusion of respondents who may have known the old recommendation of five servings per day for good health. If respondents answered outside of the range of 5–13 servings=day, they were considered not to know the recommendation for FV intake. Negative binomial regression (using the count of FV intake as outcome) and logistic regressions (for correct knowledge of FV recommendation [5–13 servings=day]). These regressions included variables that were significantly associated with the knowledge and intake outcomes at the bivariate level.
Results Characteristics of the Sample Respondents from Puerto Rico (HINTS-Puerto Rico) and from other U.S. racial=ethnic groups in HINTS 2007–2008 (U.S. non-Hispanic Whites, U.S. Blacks, and U.S. Hispanics) differed significantly (p < .05) by age, sex, education, household income, and use of health information sources (Table 1). Compared with the other racial=ethnic groups, larger percentages of the U.S. Hispanic respondents were younger (49.9% were 18–34 years old), had less than high school education (41%), and had a household income of less than $35,000 per year (51%). Meanwhile, 44.6% of respondents in Puerto Rico reported household income values below $35,000, compared with only 15.7% and 38% of Whites and Blacks, respectively, who reported incomes less than $35,000. A majority of residents from Puerto Rico and U.S. Hispanics (67.1 and 62.5%, respectively) reported never seeking information on medical or health topics, compared with 31.4 and 44.4% of U.S. non-Hispanics Whites and Blacks, respectively. Among those who did look for information, the Internet was the most common source, particularly for U.S. non-Hispanic Whites (49% of those who reported using the Internet). The groups also differed significantly in their reported intake of FV and knowledge of FV recommendation.
107
Puerto Rico respondents reported lowest intake of mean servings=day of FV (2.7 servings=day), followed by U.S. Hispanics (3.4 servings=day), U.S. non-Hispanic Blacks (3.8 servings=day) and U.S. Whites (4.3 servings=day). More than 80% of adults in Puerto Rico (89%) and U.S. Hispanic (82%) respondents did not know the recommendation for number of FV servings to maintain good health (considered as 5–13 servings=day in our analysis), compared with 62% and 77% of U.S. non-Hispanic White and Black respondents (Table 1). Negative Binomial Regression Analysis for Fruit and Vegetable Intake The first model examined FV intake, controlling for sex and education, which were significantly associated with intake at the bivariate level (see Table 2). U.S. non-Hispanic Whites (coefficient ¼ 0.40, 95% CI [0.28, 0.52]), Blacks (coefficient ¼ 0.30, 95% CI [0.15, 0.45]), and U.S. Hispanics (coefficient ¼ 0.26, 95% CI [0.12, 0.40]) reported significantly higher (p < .05) intake of FV as compared with residents from Puerto Rico. Respondents who did not seek health information reported significantly lower FV intake (coefficient ¼ 0.24, 95% CI [–0.38, 0.09]), compared with those who sought information from the health care provider. The FV intake of those who sought health information from their health care provider was not significantly different from the FV intake of those who sought information from the Internet, printed materials and other sources (p > .05). Multivariate Analysis for the Fruit and Vegetable Recommendation In a multivariable logistic regression, controlling for sex and income, U.S. non-Hispanic Whites also had significantly higher odds of knowledge of the FV recommendation (OR ¼ 3.31, 95% CI [1.98, 5.52]) compared with respondents from Puerto Rico (U.S. Hispanics and U.S. Blacks were not significantly different from Puerto Rico). Having an annual household income higher than $50,000 was associated with higher odds of knowing the FV recommendation (for those with income between $50,000 < $75,000, OR ¼ 2.01, 95% CI [1.29, 3.14]; and for those with $75,000 and more, OR ¼ 3.12, 95% [2.04, 4.76]) whereas reporting ‘‘never seeking information on health or medical topics’’ was associated with significantly lower likelihood of knowledge of the FV recommendation (OR ¼ 0.32, 95% CI [0.20, 0.52]).
Discussion Given the already reported low intake of FV among all groups in the United States (Blanck et al., 2008; Centers for Disease Control and Prevention, 2010; Colon-Lopez et al., 2013; Colon-Ramos et al., 2009; Grimm, Foltz, Blanck, & Scanlon, 2012; Thompson et al., 2005), and even lower among Hispanics in Puerto Rico (Colon-Lopez et al., 2013; Ho et al., 2006), we examined correlates with FV intake and knowledge of the FV recommendation focusing on use of source of health information. To our knowledge,
U. Colo´n-Ramos et al.
108
Table 1. Sociodemographic characteristics and knowledge of fruit and vegetable recommendations and intake using U.S. HINTS 2007–2008 data by racial=ethnic group and data from HINTS-Puerto Rico 2009 (N ¼ 4,092) U.S. HINTS
No. of participants (%) Women, % Age (years), % 18–34 35–49 50–64 65þ Education, % Less than high school High school graduate Some college College graduate Household income Less than $20,000 $20,000 < $35,000 $35,000 < $50,000 $50,000 < $75,000 $75,000 and more Refused=missing Don’t know recommendation for fruit and vegetable intake (5–13 servings=day), % Fruit and vegetable knowledge, meana Fruits, serving=dayb Vegetables, serving=dayb Health information sourcesc Never lookedd Internet Health care provider Printed materials Other
HINTS-Puerto Rico
U.S. non-Hispanic Whites
U.S. non-Hispanic Blacks
U.S. Hispanics
Puerto Ricans
2,966 (68.1%) 51.7
247 (5.7%) 61.1
298 (7.2%) 45.1
639 (14.7%) 53.7
26.4 29.0 25.5 19.1
31.7 24.4 28.1 15.9
49.9 29.9 13.6 6.6
32.2 28.2 21.6 18.0
9.5 29.2 33.8 27.5
17.0 34.3 33.0 15.7
40.5 30.4 19.8 9.2
24.0 27.8 27.9 20.3
9.3 12.4 10.0 18.5 30.5 19.4 61.9
21.0 17.0 15.4 15.5 15.4 15.8 76.5
27.3 23.7 11.6 12.9 11.7 12.8 82.9
26.6 18.0 5.8 4.0 0.9 44.6 89.8
4.3 2.1 2.2
3.3 1.9 1.9
3.3 1.9 1.5
2.7 1.6 1.1
31.4 49.2 5.0 9.4 5.0
44.4 37.0 7.3 6.8 4.6
62.5 16.3 8.8 9.5 2.9
67.1 19.1 9.8 9.6 1.0
a
‘‘How many servings of fruits and vegetables do you think the average adult should eat each day for good health?’’ ‘‘How many servings of fruits (vegetables) do you usually eat or drink each day?’’ c ‘‘The most recent time you looked for information about health or medical topics, where did you go first?’’ (only asked of respondents who reported that they had sought information about health or medical topics during the past 12 months) d ‘‘Never looked for information about health or medical topics.’’ p values
The Association Between Fruit and Vegetable Intake, Knowledge of the Recommendations, and Health Information Seeking Within Adults in the U.S. Mainland and in Puerto Rico ´ N COLO ´ N-RAMOS1, LILA J. FINNEY RUTTEN2, RICHARD P. MOSER3, VIVIAN COLO ´ N-LOPEZ4, URIYOA 5 6 ANA P. ORTIZ , and AMY LAZARUS YAROCH 1
Department of Global Health, School of Public Health and Health Services, George Washington University, Washington, District of Columbia, USA 2 Department of Health Sciences Research, Division of Epidemiology, and Population Health Science Program, Center for the Science of Health Care Delivery, Mayo Clinic, Rochester, Minnesota, USA 3 Behavioral Research Program, National Cancer Institute, Bethesda, Maryland, USA 4 Department of Health Services Administration, Graduate School of Public Health, University of Puerto Rico, San Juan, Puerto Rico 5 Department of Biostatistics and Epidemiology, Graduate School of Public Health, Medical Sciences Campus, University of Puerto Rico, San Juan, Puerto Rico 6 Gretchen Swanson Center for Nutrition, Omaha, Nebraska, USA
Health information correlates of fruit and vegetable intake and of knowledge of the fruit and vegetable recommendations were examined using bivariate and multivariate regressions with data from the 2007–2008 U.S. National Cancer Institute’s Health Information National Trends Survey in the United States and in Puerto Rico. Residents from Puerto Rico had the lowest reported fruit and vegetable intake and the lowest knowledge of the recommended servings of fruits and vegetables to maintain good health, compared with U.S. Hispanics, non-Hispanic Whites, and Blacks. Sixty-seven percent of Puerto Rican residents and 62% of U.S. Hispanics reported never seeking information on health or medical topics. In multivariate analysis, those who never sought information on health or medical topics reported significantly lower fruit and vegetable intake (coefficient ¼ 0.24; 95% CI [0.38, 0.09]), and were less likely to know the fruit and vegetable recommendations (OR ¼ 0.32; 95% CI [0.20, 0.52]), compared with those who obtained information from their health care providers. Health promotion initiatives in the United States and Puerto Rico have invested in mass media campaigns to increase consumption of and knowledge about fruit and vegetables, but populations with the lowest intake are less likely to seek information. Strategies must be multipronged to address institutional, economic, and behavioral constraints of populations who do not seek out health information from any sources.
The 2010 Dietary Guidelines urges American adults to increase fruit and vegetable (FV) intake (U.S. Department of Agriculture, 2010). Diets high in these nutrient-rich, low-calorie foods have been associated with better weight management (Rolls, Ello-Martin, & Tohill, 2004) and decreased risk of chronic diseases (U.S. Department of Agriculture, 2010). Despite educational campaigns to tout knowledge of these recommendations (Produce for Better Health & Centers for Disease Control and Prevention, n.d.), most Americans consume less than three servings a day (Thompson et al., 2005). The low consumption is possibly Address correspondence to Uriyoa´n Colo´n-Ramos, Department of Global Health, Milken Institute School of Public Health, George Washington University, 950 New Hampshire Avenue #410, Washington, DC 20052, USA. E-mail: uriyoan@ gwu.edu
the result of economic and institutional constraints, as well as behavioral and sociocultural barriers (Kraak, Story, & Swinburn, 2013; World Health Organization, 2005). As part of the latter, knowledge of the FV recommendation can be an important component of multipronged interventions to improve FV intake, because it has been associated with raised awareness and behavioral intent (National Fruit & Vegetable Alliance, 2010), and with FV intake among adults (Colon-Lopez et al., 2013; Shaikh, Yaroch, Nebeling, Yeh, & Resnicow, 2008). However, little is known about how populations with low FV consumption attain knowledge of the FV recommendation. In the U.S. mainland, there are evident dietary disparities, with U.S. Hispanics reporting higher consumption compared with U.S. non-Hispanic Whites and Blacks (Blanck, Gillespie, Kimmons, Seymour, & Serdula, 2008; ColonRamos et al., 2009; Thompson et al., 2005). In Puerto Rico,
U. Colo´n-Ramos et al.
106 a commonwealth of the United States where 99% of the residents are Hispanic (U.S. Census Bureau, 2010), only 13.7% of the population consumes at least five servings of FV a day, compared with 24.3% of the general U.S. population (Centers for Disease Control and Prevention, 2010). Understanding health information correlates of these dietary disparities and of knowledge of the FV recommendation among racial=ethnic groups in the U.S. mainland and in Puerto Rico will shed light on how to best promote FV intake in these diverse populations. We used data collected via the U.S. National Cancer Institute’s Health Information National Trends Survey in Puerto Rico (HINTS-Puerto Rico) and U.S. mainland (HINTS 2007–2008) (Tortolero-Luna et al., 2010) to elucidate potential health information sources related to FV consumption in Puerto Rico and in the U.S. mainland (non-Hispanic Whites, Blacks, and mainland Hispanics).
Method HINTS is a national survey designed to track population trends in cancer-relevant behavior, knowledge, and attitudes and assess the American public’s use of and experiences with health information (National Cancer Institute, n.d.). Data were collected from HINTS 2007–2008 in the U.S. mainland, and a special data collection effort that used the HINTS 2007–2008 items in Puerto Rico (HINTS-Puerto Rico) in 2009 (Tortolero-Luna et al., 2010). The HINTS 2007–2008 instrument includes previous iterations of HINTS questions (about 50% of the items) and new questions that were developed and reviewed by working groups at the National Cancer Institute. The instrument underwent three rounds of cognitive tests. The instrument was also translated into Spanish by a native Spanish speaker with extensive experience in survey translation. The instrument was reviewed, edited, and pilot-tested for the island population by the HINTS Puerto Rico team. Data Collection and Response Rates HINTS 2007–2008 employed a mixed mode, dual-frame design (n ¼ 7,674) including list-assisted, random-digit dial computer-assisted telephone interview (n ¼ 4,092) and a self-administered mailed survey that used a postal frame of U.S. addresses (n ¼ 3,582). Using the same instrument (translated to Spanish), HINTS-Puerto Rico was administered solely as a random digit dialing survey (n ¼ 639); therefore, to keep mode consistent, our analyses were restricted to the random digit dialing sample for the U.S. mainland. In the U.S. mainland, the random digit dialing survey was conducted from January to April 2008. The random digit dialing household screener response rate was 42.4%, and the response rate for extended interview was 57.2%, with an overall response rate of 24.2%. In Puerto Rico, a stratified sampling frame was used representing the eight geographic regions of the Puerto Rico Department of Health. The survey in Puerto Rico was conducted by bilingual interviewers (Spanish and English) between April and June 2009, with a response rate of 76.3% and a total of 639 interviews
completed. Greater details on the HINTS 2007–2008 sample and sampling design, and HINTS-Puerto Rico have been published elsewhere (Cantor et al., 2009; Tortolero-Luna et al., 2010). Study Variables Sociodemographic characteristics—sex, education, age, and household income—were collected. Diet-related variables and health information variables were assessed. Diet-Related Variables Two questions were used to asses FV intake: 1. ‘‘How many servings of fruits do you usually eat or drink each day? Think of a serving as being about 1 medium piece, or 1=2 cup of fruit, or 3=4 of cup of fruit juice.’’ 2. ‘‘How many servings of vegetables do you usually eat or drink each day? Think of a serving as being about 1 cup of raw leafy vegetables, 1=2 cup of other cooked or raw vegetables, or 3=4 cup of vegetable juice.’’ Both questions were open-ended, and participants were asked to write in a number for servings each day. Validity for these questions was established for the U.S. mainland population (Yaroch et al., 2012). Knowledge of FV recommendation was assessed by asking, ‘‘How many servings of fruits and vegetables do you think the average adult should eat each day for good health?’’ Responses were open-ended and calculated as dichotomous (those who responded anywhere between 5–13 servings=day vs. those who responded outside of that range) and continuous. Health Information Variables Participants were asked the following: 1. ‘‘Have you ever looked for information about health or medical topics from any source?’’ (Y=N) 2. ‘‘The most recent time you looked for information about health or medical topics, where did you go first?’’ (Twelve choices were provided with the option to write in ‘other’ choices) Participant responses were then grouped into the following: Internet, doctor or health care provider, printed materials (library, magazines, newspapers, books, brochures, pamphlets), and other (cancer organization, family, friend or coworker, telephone information number, or alternative practitioner). Statistical Analyses We used SAS-callable SUDAAN version 9.0 to account for the complex survey design of HINTS 2007–2008 and HINTS-Puerto Rico. All analyses used sample weights to obtain population-level estimates of the populations in the mainland U.S. and Puerto Rico, respectively. Weighted frequencies were calculated for Puerto Rico and by race= ethnicity in the United States (U.S. non-Hispanic Whites, U.S. non-Hispanic Blacks, and U.S. Hispanics). The majority of respondents in HINTS-Puerto Rico (99%) were of
Health Information Correlates of Diet in Puerto Rico and United States Hispanic=Latino origin. Also, almost half of the respondents in HINTS-Puerto Rico (44.6%) refused to report income or had missing values, so we report them as refused=missing. We first summarized weighted frequencies for each of the variables stratified by race=ethnicity, and calculated a p value derived from a chi-square statistic (for categorical variables) or Wald F statistic (for continuous values) to assess differences across distribution of sociodemographic and health information characteristics by racial=ethnic groups. These tests were also used to identify potential associations for each variable of interest with FV intake or knowledge of FV recommendation (the latter results are not shown). Median and means of FV intake and of responses to the number of servings needed for good health were examined. Kurtosis and skewness values were examined to assess assumption of normal distribution for both variables. We also dichotomized the responses for the knowledge variable (knowledge on the number of servings needed for good health) into two categories: 5–13 servings=day and anything outside of that range. We used the more comprehensive range of 5–13 servings=day to allow for inclusion of respondents who may have known the old recommendation of five servings per day for good health. If respondents answered outside of the range of 5–13 servings=day, they were considered not to know the recommendation for FV intake. Negative binomial regression (using the count of FV intake as outcome) and logistic regressions (for correct knowledge of FV recommendation [5–13 servings=day]). These regressions included variables that were significantly associated with the knowledge and intake outcomes at the bivariate level.
Results Characteristics of the Sample Respondents from Puerto Rico (HINTS-Puerto Rico) and from other U.S. racial=ethnic groups in HINTS 2007–2008 (U.S. non-Hispanic Whites, U.S. Blacks, and U.S. Hispanics) differed significantly (p < .05) by age, sex, education, household income, and use of health information sources (Table 1). Compared with the other racial=ethnic groups, larger percentages of the U.S. Hispanic respondents were younger (49.9% were 18–34 years old), had less than high school education (41%), and had a household income of less than $35,000 per year (51%). Meanwhile, 44.6% of respondents in Puerto Rico reported household income values below $35,000, compared with only 15.7% and 38% of Whites and Blacks, respectively, who reported incomes less than $35,000. A majority of residents from Puerto Rico and U.S. Hispanics (67.1 and 62.5%, respectively) reported never seeking information on medical or health topics, compared with 31.4 and 44.4% of U.S. non-Hispanics Whites and Blacks, respectively. Among those who did look for information, the Internet was the most common source, particularly for U.S. non-Hispanic Whites (49% of those who reported using the Internet). The groups also differed significantly in their reported intake of FV and knowledge of FV recommendation.
107
Puerto Rico respondents reported lowest intake of mean servings=day of FV (2.7 servings=day), followed by U.S. Hispanics (3.4 servings=day), U.S. non-Hispanic Blacks (3.8 servings=day) and U.S. Whites (4.3 servings=day). More than 80% of adults in Puerto Rico (89%) and U.S. Hispanic (82%) respondents did not know the recommendation for number of FV servings to maintain good health (considered as 5–13 servings=day in our analysis), compared with 62% and 77% of U.S. non-Hispanic White and Black respondents (Table 1). Negative Binomial Regression Analysis for Fruit and Vegetable Intake The first model examined FV intake, controlling for sex and education, which were significantly associated with intake at the bivariate level (see Table 2). U.S. non-Hispanic Whites (coefficient ¼ 0.40, 95% CI [0.28, 0.52]), Blacks (coefficient ¼ 0.30, 95% CI [0.15, 0.45]), and U.S. Hispanics (coefficient ¼ 0.26, 95% CI [0.12, 0.40]) reported significantly higher (p < .05) intake of FV as compared with residents from Puerto Rico. Respondents who did not seek health information reported significantly lower FV intake (coefficient ¼ 0.24, 95% CI [–0.38, 0.09]), compared with those who sought information from the health care provider. The FV intake of those who sought health information from their health care provider was not significantly different from the FV intake of those who sought information from the Internet, printed materials and other sources (p > .05). Multivariate Analysis for the Fruit and Vegetable Recommendation In a multivariable logistic regression, controlling for sex and income, U.S. non-Hispanic Whites also had significantly higher odds of knowledge of the FV recommendation (OR ¼ 3.31, 95% CI [1.98, 5.52]) compared with respondents from Puerto Rico (U.S. Hispanics and U.S. Blacks were not significantly different from Puerto Rico). Having an annual household income higher than $50,000 was associated with higher odds of knowing the FV recommendation (for those with income between $50,000 < $75,000, OR ¼ 2.01, 95% CI [1.29, 3.14]; and for those with $75,000 and more, OR ¼ 3.12, 95% [2.04, 4.76]) whereas reporting ‘‘never seeking information on health or medical topics’’ was associated with significantly lower likelihood of knowledge of the FV recommendation (OR ¼ 0.32, 95% CI [0.20, 0.52]).
Discussion Given the already reported low intake of FV among all groups in the United States (Blanck et al., 2008; Centers for Disease Control and Prevention, 2010; Colon-Lopez et al., 2013; Colon-Ramos et al., 2009; Grimm, Foltz, Blanck, & Scanlon, 2012; Thompson et al., 2005), and even lower among Hispanics in Puerto Rico (Colon-Lopez et al., 2013; Ho et al., 2006), we examined correlates with FV intake and knowledge of the FV recommendation focusing on use of source of health information. To our knowledge,
U. Colo´n-Ramos et al.
108
Table 1. Sociodemographic characteristics and knowledge of fruit and vegetable recommendations and intake using U.S. HINTS 2007–2008 data by racial=ethnic group and data from HINTS-Puerto Rico 2009 (N ¼ 4,092) U.S. HINTS
No. of participants (%) Women, % Age (years), % 18–34 35–49 50–64 65þ Education, % Less than high school High school graduate Some college College graduate Household income Less than $20,000 $20,000 < $35,000 $35,000 < $50,000 $50,000 < $75,000 $75,000 and more Refused=missing Don’t know recommendation for fruit and vegetable intake (5–13 servings=day), % Fruit and vegetable knowledge, meana Fruits, serving=dayb Vegetables, serving=dayb Health information sourcesc Never lookedd Internet Health care provider Printed materials Other
HINTS-Puerto Rico
U.S. non-Hispanic Whites
U.S. non-Hispanic Blacks
U.S. Hispanics
Puerto Ricans
2,966 (68.1%) 51.7
247 (5.7%) 61.1
298 (7.2%) 45.1
639 (14.7%) 53.7
26.4 29.0 25.5 19.1
31.7 24.4 28.1 15.9
49.9 29.9 13.6 6.6
32.2 28.2 21.6 18.0
9.5 29.2 33.8 27.5
17.0 34.3 33.0 15.7
40.5 30.4 19.8 9.2
24.0 27.8 27.9 20.3
9.3 12.4 10.0 18.5 30.5 19.4 61.9
21.0 17.0 15.4 15.5 15.4 15.8 76.5
27.3 23.7 11.6 12.9 11.7 12.8 82.9
26.6 18.0 5.8 4.0 0.9 44.6 89.8
4.3 2.1 2.2
3.3 1.9 1.9
3.3 1.9 1.5
2.7 1.6 1.1
31.4 49.2 5.0 9.4 5.0
44.4 37.0 7.3 6.8 4.6
62.5 16.3 8.8 9.5 2.9
67.1 19.1 9.8 9.6 1.0
a
‘‘How many servings of fruits and vegetables do you think the average adult should eat each day for good health?’’ ‘‘How many servings of fruits (vegetables) do you usually eat or drink each day?’’ c ‘‘The most recent time you looked for information about health or medical topics, where did you go first?’’ (only asked of respondents who reported that they had sought information about health or medical topics during the past 12 months) d ‘‘Never looked for information about health or medical topics.’’ p values
