The Association of Understanding of Medical Statistics with Health Information Seeking and Health Provider Interaction in a National Sample of Young Adults

Abstract

Numeracy, or, “the ability to use and understand numbers in daily life” is a critical component of health literacy. However, little research has focused on numeracy in young adults (ages 18–29). We used a national sample to examine how health-information seeking, trust in sources, and interactions with health care providers differ for young adults with lower and higher numeracy. We included respondents ages 18 to 29 (n = 661) from the latest administration (2008) of the Health Information National Trends Survey (HINTS). There were no significant differences between those with lower and higher numeracy for most sociodemographic variables, nor did numeracy predict trust in health information sources. However, there were several differences for health-information seeking and health-provider interactions. Those with lower numeracy were significantly more likely to say their most recent search took a lot of effort (46% vs. 24%, p = .0008) and was frustrating (45% vs. 22%, p = .0038). Those in the lower numeracy group also reported more negative interactions with health providers, including feeling less able to rely on their provider (62% vs. 86%, p < .0001), and less likely to say their provider made sure they understood information (70% vs. 88%, p = .0001) and helped with any uncertainty (51% vs. 75%, p < .0001), even when adjusting for other variables. Our data suggest that limited comfort with numbers and statistics can influence a variety of health-related factors for young adults. More research is needed to understand how health literacy skills—including numeracy—influence health-information seeking, patient-provider relationships, and health outcomes, for young adults.

Health literacy has been defined in multiple ways. A recent definition states “Health literacy occurs when the skills and ability of those requiring health information and services are aligned with the demand and complexity of information and services” (Parker & Ratzan, 2010). Numeracy, or, “the ability to use and understand numbers in daily life” (Rothman et al., 2006), is often considered a component of health literacy (Nielsen-Bohlman, Panzer, & Kindig, 2004), because the ability to understand items such as nutrition labels or medication directions often requires a working knowledge of numbers and simple math skills. Further, numeracy skills, such as addition, subtraction, and multiplication, are critical in helping people understand information about risk and likelihood.

The specific skill of numeracy has been studied less often than the broader concept of health literacy, but the studies that exist suggest that low numeracy skills can have a negative impact on health. One review reports that “low numeracy distorts perceptions of risks and benefits of screening, reduces medication compliance, impedes access to treatments, impairs risk communication … [and] appears to adversely affect medical outcomes” (Reyna, Nelson, Han, & Dieckmann, 2009). Low numeracy can also limit one's ability to self-manage illness (Montori & Rothman, 2005), to be an active participant in decision-making about one's health (Galesic & Garcia-Retamero, 2011; Montori & Rothman, 2005), and to assess risk and benefit information (Peters, Hibbard, Slovic, & Dieckmann, 2007). Studies also show that low numeracy skills are associated with lower self-efficacy for glycemic control (Osborn, Cavanaugh, Wallston, & Rothman, 2010) and fewer self-management behaviors (Cavanaugh et al., 2008) among patients with diabetes. In the areas of cancer screening, low numeracy has been related to decreased understanding of mammogram benefits (Schwartz, Woloshin, Black, & Welch, 1997), treatment (Lipkus, Peters, Kimmick, Liotcheva, & Marcom, 2010), and genetic counseling information (Portnoy, Roter, & Erby, 2010) for breast cancer and lower rates of colorectal screening (Ciampa, Osborn, Peterson, & Rothman, 2010). Low numeracy skills have also been associated with problems in evaluating medication side effects (Gardner, McMillan, Raynor, Woolf, & Knapp, 2011), understanding nutrition labels (Rothman et al., 2006), having a higher body mass index (BMI; Huizinga, Beech, Cavanaugh, Elasy, & Rothman, 2008), and both poor communication (Ciampa, Osborn, Peterson, & Rothman, 2010) and interaction (Smith, Wolf, & von Wagner, 2010) with providers.

Although many studies include participants aged 18 years and older, the focus has mostly been on older adults. The average age of participants identified in studies mentioned above ranges from 37 to 68 (Cavanaugh et al., 2008; Ciampa et al., 2010; Estrada, Martin-Hryniewicz, Peek, Collins, & Byrd, 2004; Galesic & Garcia-Retamero, 2011; Gardner et al., 2011; Huizinga et al., 2008; Lipkus et al., 2010; McAuliffe, DiFranceisco, & Reed, 2010; Osborn et al., 2010; Portnoy et al., 2010; Rothman et al., 2006; Schwartz et al., 1997). Thus, although knowledge about numeracy has been growing, research has not focused on young adults; therefore, little is known about prevalence rates of low numeracy for this population, or how low numeracy affects this age group. One estimate from the 2003 National Assessment of Adult Literacy found that health literacy skills increase as people age, and then begin to decline at around age 40, suggesting that young adults are in a stage of still acquiring skills (Kutner, Greenberg, Jin, & Paulsen, 2006). In a study of medical students that focused on young adults (median age 24), researchers found that 77% could answer all three numeracy questions correctly, and lower numeracy limited the ability of students to evaluate quantitative data, suggesting that numeracy issues can even exist for young adults with a high level of education (Sheridan & Pignone, 2002). Another study found that people in the 18 to 34 age group were more likely to report statistical confidence than adults in older age groups (Smith et al., 2010).

Young adults, often considered ages 18 to 29 (Rindfuss, 1991), compose an important age group to examine in studies of health literacy and numeracy. As adolescents transition into adulthood, not only are they still acquiring skills, but they are also at a point in their lives when they are often living independently for the first time. Between the ages of 18 and 24, only 50% of young adults remain living with parents, and by the ages of 25 to 29, only 15% remain in their parents' home (Rumbaut & Komaie, 2007). There is much diversity in living situations; although some young adults go to college, others join the workforce (Rumbaut & Komaie, 2007), and there are multiple pathways to independence among this age group (Park, Mulye, Adams, Brindis, & Irwin, 2006; Rindfuss, 1991). This variation in transitions suggests that multiple approaches to addressing health care needs are advised, because some young adults may be more independent, have different skill sets, and have varying levels of access to health insurance and health services.

Regardless of the pathway taken, people in this age group are making decisions on their own, with legal responsibility for their health care; typically they are no longer under regular supervision from parents or guardians, which suggests that young adulthood is a prime time to ensure that health literacy and numeracy skills are developed and that young adults are able to function within the health care system at a high level. In fact, many in this age group are becoming parents themselves; 36% of women ages 18 to 29 in 2010 had a child (Wang & Taylor, 2011), and since they may be managing the health care of their children as well, it is crucial to understand any limitations that may exist.

At the same time young adults are beginning to live independently, they are also facing significant challenges to their health. Park and colleagues report that people ages 18 to 24 have twice the mortality rate as teens ages 12 to 17, with males having a higher mortality rate than females (Park et al., 2006). Park and colleagues also report that the leading cause of mortality for both males and females is unintentional injury, and that violence, reproductive health, mental health, and tobacco, alcohol, and other substance use are the leading health issues for this age group. While dealing with these health issues, this age group is also faced with a lack of health insurance. In a recent study, only 62% reported having health insurance, a lower percentage than other age groups (Taylor & Keeter, 2010).

Given the increasing independence of young adults and the health issues this segment of the population faces, it is important to ascertain the rates of low health literacy and numeracy among young adults, and to understand what issues young adults with low health literacy and numeracy skills may have with respect to making health decisions and interacting with the health care system. In addition, those aged 18 to 29 in recent years are part of the Millenial generation, which is characterized by widespread Internet (90%) and cell phone (94%) use compared with other generations, and increased use of the Internet to get information (Taylor & Keeter, 2010). However, one recent study of low-income adults found that those with low numeracy were less likely to have computers and cell phones (Jensen, King, Davis, & Guntzviller, 2010), and limitations in education and literacy skills have been found to influence the ability of individuals to engage in online health-information seeking (Cotten & Gupta, 2004), suggesting that not all members of this generation have the instant access to information that is expected. Health information seeking is a complex behavior that includes not only the ability to access and understand information, but also the motivation to search for information, as well as the knowledge of how to find relevant information, how to exclude sources that are not credible, and how to apply the information found in order to make decisions. Thus, although this generation might have expanded options for accessing health information, many of the same barriers to understanding and making use of the information remain.

There is a need to improve understanding of numeracy issues for young adults, as described earlier; however, most studies examining rates of numeracy and health outcomes related to numeracy have focused on older adults, and only one study we identified examined data from a large, national U.S. sample (Ciampa et al., 2010). Thus, our study uses a nationally representative U.S. sample to answer questions regarding numeracy and young adults in order to address the following aims:

	Identify how trust in health information sources varies among young adults with lower and higher numeracy skills.
	Examine how health-information-seeking experiences differ for young adults with lower and higher numeracy skills.
	Assess how numeracy is associated with interactions with health care providers.

Methods

The Health Information National Trends Survey (HINTS) is a national probability sample of the adult population developed by the National Cancer Institute. Details of the survey have been published elsewhere (Cantor et al., 2009; Nelson et al., 2004). The last administration of the HINTS survey was conducted between January and May 2008 (although the administration being used is referred to as HINTS 2007) via phone (n = 4,092) and mail (3,582). The survey questionnaire and data set are available online (National Cancer Institute, 2011). We included both English- and Spanish-language respondents, and included all respondents ages 18 to 29 (n = 661) without a missing value for the numeracy question. Of these 661 respondents, 393 completed the survey via mail and 268 completed the survey via telephone. Of those who were surveyed by telephone, 35 respondents completed the interview in Spanish. This study was classified as exempt by the University at Albany Institutional Review Board.

Measures

Given that this was a secondary data analysis, we chose our measure of numeracy based on what was available. We used the question “In general, how easy or hard do you find it to understand medical statistics?” as a proxy measure for numeracy (and refer to this measure as numeracy throughout the paper). This question represents confidence in using statistics as opposed to actual ability (Woloshin, Schwartz, & Welch, 2005), and was called statistical confidence in one study (Smith et al., 2010), but we believed it was an appropriate proxy measure for numeracy, given that numeracy related to medical settings is heavily dependent on statistics and has been referred to as numeracy in other research (Ciampa et al., 2010). We placed 196 respondents (30%) who replied “Hard”/“Very hard” in the lower numeracy group, and 465 respondents (70%) who replied “Easy”/“Very easy” in the higher numeracy group. We found this measure to be strongly associated with the other three numeracy questions asked on the survey (Appendix), and used this single item as opposed to creating a scale of the various items asked, because we thought it was a reasonable proxy measure to use and because of its high correlation with the other items.

We compared numeracy group with several variables related to sociodemographics (Table 1). The categories used for education were (a) did not complete high school, (b) high school graduate, (c) some college, and (d) college graduate or higher. Household income as asked was divided into five categories: (a) missing, (b) <$20,000, (c) $20,000–$49,999, (d) $50,000–$74,999, and (e) $75,000 or higher. We examined whether the respondent was employed or not, married or not, was born in the U.S. or not, and whether they were male or female. Race/ethnicity data used existing variables that combined race and ethnicity to create the resulting groups: White non-Hispanic, African American non-Hispanic, Hispanic of any race, and Other. We also examined measures of health status, whether or not there was a regular health provider, and health insurance status. The health status variable had three categories: (a) excellent/very good, (b) good, and (c) fair/poor.

Table 1. Demographic, health status, and health care access characteristics for people with lower and higher numeracy, weighted (n = 661)

	% of sample	Lower numeracy	Higher numeracy	p value
Race/Ethnicity				.0458
White (non-Hispanic)	59%	51%	63%
African American (non-Hispanic)	12%	13%	11%
Other (non-Hispanic)	7%	5%	8%
Hispanic	22%	31%	18%
Education				.0719
Less than high school	13%	21%	10%
High school grad	30%	30%	30%
Some college	38%	36%	40%
College grad or higher	18%	13%	20%
Income				.2107
Missing	12%	14%	10%
<$20,000	23%	31%	20%
$20,000–$49,999	30%	26%	31%
$50,000–$74,999	17%	16%	17%
$75,000 or higher	19%	14%	21%
Born in the United States	84%	81%	86%	.3535
Female	49%	46%	50%	.5868
Employed	53%	54%	53%	.8702
Married	23%	26%	22%	.3777
Health status				.0017
Excellent/very good	50%	35%	57%
Good	36%	46%	31%
Fair/poor	14%	19%	12%
Health care access
Have regular provider	49%	46%	51%	.4112
Have health insurance	71%	63%	74%	.0921

We also used numeracy group as an independent variable in models with trust in health information sources, health-information-seeking practices, and health provider communication as dependent variables. Trust in information sources (see Table 2) was measured by asking the respondents separate questions for each health information source: “In general, how much would you trust information about health or medical topics from … (newspapers or magazines, radio, television, Internet, family or friends, doctor or other health care professional?).” Responses used a 4-point scale ranging from “A lot” (1) to “Not at all” (4). These were reverse coded for analysis, and we created an average score. Questions related to health-information seeking are listed in Table 3. These questions included items about whether the respondent ever searched for health or medical information, and, for their most recent search, where they went to get information and their experience with searching for information (e.g., it took a lot of effort).

Table 2. Comparison of mean scores for trust in health information sources for people with lower and higher numeracy, weighted (n = 661)

	Lower numeracy	Higher numeracy	p value
TV	2.25 ± .08	2.32 ± .04	.4153
Radio	2.06 ± .09	2.16 ± .05	.3317
Internet	2.77 ± .10	2.84 ± .05	.5500
News/mags	2.36 ± .09	2.46 ± .04	.2783
Family/friends	2.88 ± .10	2.78 ± .05	.3503
Doctor	3.54 ± .09	3.71 ± .04	.0825

Table 3. Health information seeking characteristics for people with lower and higher numeracy, weighted (n = 661)

	% of sample	Lower numeracy	Higher numeracy	p value, un-adjusted	p value, adjusted*
Ever looked for information on health or medical topics?	64%	61%	65%	.5014	.715
Where did you go first the most recent time?				.2513	NA
Internet	75%	75%	75%
Books/magazines/news	8%	5%	9%
Health provider	7%	6%	7%
Friends/family	4%	8%	3%
Other	6%	7%	6%
Agree that most recent search:
took a lot of effort	31%	46%	24%	.0008	.037
was frustrating	29%	45%	22%	.0038	.026
concerned about quality	54%	64%	50%	.0299	.085

Questions about health provider communication (Table 4) focused on functions of patient-centered communication (Arora, 2009) and asked items such as whether the provider always or usually made sure the respondent understood information given during the past 12 months. We combined “Always” and “Usually” into one response category, and “Sometimes” and “Never” into another category. We also examined questions concerning self-efficacy, which asked about confidence in taking care of one's health and confidence in getting health information if it were necessary (Moldovan & Heald, 2009).

Table 4. Health provider experience for people with lower and higher numeracy, weighted (n = 661)

	% of sample	Lower numeracy	Higher numeracy	p value, un-adjusted	p value, adjusted*
Provider communication
During past 12 months, health provider always or usually:
Gave chance to ask questions	84%	79%	86%	.1808	.487
Addressed feelings	73%	60%	79%	.0009	.005
Involved in decisions	71%	55%	77%	.0001	.004
Made sure understood info	83%	70%	88%	.0001	.003
Helped with any uncertainty	67%	51%	75%	<.0001	<.0001
Other provider measures
During past 12 months, felt could rely on provider to take care of health needs	79%	62%	86%	<.0001	<.0001
Quality of care past 12 months				.0003	.004
Excellent/very good	65%	48%	72%
Good	23%	27%	21%
Fair/poor	13%	25%	7%
Avoid doctor	40%	49%	35%	.0246	.06
Self-efficacy
Confident can take care of health				<.0001	.003
Very confident	70%	52%	78%
Somewhat confident	28%	44%	21%
Not at all confident	2%	4%	1%
Confident could get health information if needed to				<.0001	.001
Very confident	63%	45%	70%
Somewhat confident	32%	43%	28%
Not at all confident	5%	13%	2%

Analysis

We used Pearson's coefficient to compare weighted differences between continuous variables and chi-square tests to compare responses among categorical variables. We used multivariate logistic regression and multivariate ordinal logistic regression to examine the relationship between numeracy and outcome variables, adjusting for mode effects as well as variables that were significant or close to significant in bivariate analyses presented in Table 1. Analyses were conducted using STATA SE version 11 (StataCorp, College Station, Texas), and survey (svy) commands were used to incorporate combined (phone and mail) sampling weights (StataCorp, 2011).

Results

The average age of the sample of young adults was 23.3. There were no significant differences between those with lower (30% of sample) and higher (70% of sample) numeracy for education level, gender, born in the U.S., employment status, marital status, and household income; and race/ethnicity (p = .0458) was borderline significant (Table 1). Although numeracy was associated with health status (p = .0017), it was not related to having a regular health provider or health insurance status. It also appears that numeracy did not predict opinions about trust in various health information sources (Table 2), although, of note, the source that came closest to achieving statistical significance was doctor (p = .0825), with people in the lower numeracy group reporting less trust in their doctor than people in the higher numeracy group.

Although there were no significant differences with respect to demographics or trust in information sources, there were numerous differences with regard to health-information seeking and health-provider interactions. Table 3 presents data examining differences in health-information seeking for those with lower and higher numeracy. Both groups report looking for health information in almost equal numbers (61% vs. 65%), and report using similar information sources, with an overwhelming number reporting the Internet as the first place they went to on their most recent search for health information (75% both groups). However, those with lower numeracy were significantly more likely to say that their most recent search took a lot of effort (p = .001), was frustrating (p = .004), and yielded information that led to concerns about quality (p = .03); search effort and frustration continued to be significant even in adjusted models.

When asked about experiences with health care providers, those in the lower numeracy group seemed to be less satisfied with interactions, as shown in Table 4. People in the lower numeracy group were more likely to rate their quality of care as fair/poor than those in the higher numeracy group (p = .0003). They were also less likely to feel that they could rely on their provider (p < .0001) and were less likely to say that their provider made efforts to make sure they understood the information (p = .0001) and helped with any uncertainty (p < .0001). Finally, those in the lower numeracy group also appeared to have lower self-efficacy than those in the higher numeracy group. They were significantly less likely to be confident that they could take care of their health (p < .0001) and obtain health information if necessary (p < .0001). Numeracy remained significant in adjusted models with the exception of avoiding doctors (Table 4), and was the only significant predictor in many of the models.

We ran the same set of models separately for adults ages 30 and over (n = 6,898, average age = 52.2) to determine whether the results would be the same. Among people ages 30 and over, 40% had lower numeracy and 60% had higher numeracy. Unlike our observations of young adults, there were statistically significant differences between the lower and higher numeracy group for all demographic variables except for gender. The most striking differences suggest that those in the lower numeracy group were more likely to have less education and lower income, and to be unemployed. We also observed statistically significant differences between lower and higher numeracy groups for trust in information sources (with the exception of family), and numeracy was a significant predictor in multivariate models for all of the variables concerning health-information seeking and interaction with health providers. In contrast to models with young adults, health status and health insurance often appeared as significant factors related to health-information seeking and health-provider interactions, in addition to numeracy.

Discussion

Research has indicated that numeracy skills are an issue for adults of all ages, but no study has specifically reported on numeracy for a nationally representative young-adult sample. Our data suggest that numeracy does impact young adults with regard to health-information seeking and health-provider communication. In our sample, although younger adults with low numeracy skills did not differ by sociodemographic characteristics from those with high numeracy, this was not true for older adults, suggesting that traditional variables, such as education or socioeconomic status, may not be effective for identifying low numerate groups in this age cohort. This could be because some young adults have not completed their education, and income differences may be less pronounced early in the working years. Although numeracy appeared to relate to trust in information sources for older adults, this was not true for young adults. We did find that numeracy was associated with health-information seeking and health-provider interaction for both younger and older adults. However, it appeared that other factors, such as health status, also were important factors for older adults. Our finding regarding numeracy and health-provider interaction is similar to studies that also used HINTS data (Ciampa et al., 2010; Smith et al., 2010).

Public health professionals, health care providers, and others who are involved with providing health information should consider that young adults might have issues with understanding numerical information and may need to be given information in a way that accounts for this. Because of the significant health risks facing this age group, it is crucial to ensure that the information made available is easy to understand, and that providers and practitioners communicate clearly with young adults. In addition, because young adults often have children, and limited numeracy skills can affect actions such as whether parents give their children the correct dosage of medications (Sanders, Federico, Klass, Abrams, & Dreyer, 2009), the need to address low numeracy in young adults is even more important.

Our results found that young adults report issues with obtaining health information when they have a limited comfort level using numbers and statistics. Because people who are less able to understand numbers and statistics have a more difficult time with health-information seeking, it will perhaps be useful to consider ways to provide people identified as having low numeracy with skills that could improve the information seeking process. One area of focus for future interventions could be to help people become more comfortable with numbers and statistics in order to increase numeracy skills among young adults. It is equally important to develop interventions that can provide young adults with the skills to seek information on their own.

People might assume that, because the Millenial generation is technologically advanced and has high Internet use, they have instant access to information that could enhance their interactions with the health care system. Although this may be true, it is important to consider how information is typically accessed. A recent study showed that people ages 18 to 29 were more likely to use cell phones to look for health information compared to those in older age groups, suggesting that health-information-seeking patterns are perhaps unique for the Millenial generation (Fox, 2011). Thus, health providers should recognize that the media being used by young adults to obtain information may not be the same as those used by older adults, and should ensure that such information is easily accessible by mobile technology. Anyone who develops health-information materials or websites should also note that information designed to be read on a computer screen might not translate easily to a much smaller mobile device.

It is also important to consider that having instant access to information does not necessarily mean that someone can understand the information they obtain, apply it to make decisions, or assess the accuracy of the information. A study of young adults found that Internet activities differed based on self-report of Internet skills and education level (Hargittai & Hinnant, 2008). According to other research, most websites with health-related information require high reading levels (Berland et al., 2001), and statistical literacy is needed in order to use the numeric information that people are able to so easily access (Wallman, 2010).

Given this, attention should be paid to how health information is provided and how to improve the skills of young adults accessing information. In addition to enhancing numeracy skills—as well as broader health literacy skills—improving media literacy skills, which include the ability to assess credibility of information found online, is important for the Millenial generation (Considine, 2008). Also important is ehealth literacy, which is “the ability to seek, find, understand, and appraise health information from electronic sources and apply the knowledge gained to addressing or solving a health problem” (Norman & Skinner, 2006).

Young adults are beginning to navigate their way within the health system as independent users of health services. The knowledge that people in their 20s from all racial, ethnic, and socioeconomic status groups may have limited literacy and numeracy skills should inform health providers of the need to account for patient health literacy and numeracy skills when interacting with young adult patients and explaining information to them. Techniques such as using clear language, using diagrams or pictures, asking patients to recall information and “teach back” what they were told, facilitating decision making and accessing health care services, and encouraging patients to ask questions can all provide assistance to a patient with low health literacy and numeracy skills (Epstein & Street, 2007). Specific to numeracy, useful strategies include reducing the necessity for patients to perform calculations and providing only the most significant and necessary amount of information (Peters et al., 2007).

There is no single, universally agreed upon definition and measurement of numeracy; however, there are several tools that have been developed that can help provide an assessment of numeracy skills. Such tools include the Test of Functional Health Literacy in Adults (TOFHLA) (Nielsen-Bohlman et al., 2004), the Newest Vital Sign (NVS; Weiss et al., 2005), and a scale related to medical statistics as used in the HINTS survey (Woloshin et al., 2005), among others (McCormack et al., 2010). Providers can use these tools to help identify patients who need extra assistance to understand information, and researchers can utilize these tools to further study health literacy, including numeracy, as an issue for young adults.

Our study represents an important first step in providing information about how numeracy skills affect young adults with health-information seeking and with interactions with health care providers, but there are limitations of the research. One limitation concerns our measurement of numeracy. The item used is part of a larger three-item scale (Woloshin et al., 2005), but this was the only item from the scale included in the HINTS survey. Although it was a single item measure, other research has found that a single item can be accurate in predicting health literacy skills (Chew et al., 2008; Morris, MacLean, Chew, & Littenberg, 2006; Sarkar, Schillinger, López, & Sudore, 2011). In addition, the item used does not fully reflect a person's ability to understand numbers and perform calculations. However, the use of this question as a proxy measure allowed for an initial investigation of the issues. Also, the data do not include a measure of functional literacy. It may be that some people with low numeracy may have adequate functional literacy, and are better able to navigate through health information and the health system than those with both low numeracy and low health literacy skills. Given these issues, it is important to note that the findings are preliminary, and these questions should be addressed in future research with more robust measures of health literacy and numeracy.

Park and colleagues argue that an agenda must be created to address health issues for young adults (Park et al., 2006), and we believe that agenda must include numeracy as well as health literacy. The field needs to conduct more research to explore how numeracy skills, as well as general health literacy abilities, influence health-information seeking, interactions with the health system, and health outcomes for young adults, and whether such relationships vary among specific subgroups of the young-adult population with lower health literacy and numeracy skills. It is also crucial that we work to develop interventions to enhance numeracy skills for young adults, because we could find no evidence of interventions that have been specifically designed to address numeracy skills in the young-adult population. Interventions can be developed to specifically target young adults in school or in the workforce. By creating programs that are tailored to this age group, professionals can ensure that both the content and administration of such programs is appropriate. By intervening at the stage when youth are transitioning to an independent lifestyle, we can identify those at risk and enhance their ability to seek and understand health information, as well as ease their interactions with the health care system, in order to maximize their potential for health and well-being.

Notes

*The association of numeracy with these variables is adjusted for race/ethnicity, education level, health status, and mode of survey administration (phone vs. mail).

*The association of numeracy with these variables is adjusted for race/ethnicity, education level, health status, and mode of survey administration (phone vs. mail).

Appendix: Comparison of Numeracy Variables with Other Numeracy Measures (n = 661)

Table

	Lower numeracy	Higher numeracy	p value
When people tell you the chance of something happening, do you prefer they tell you with:			<.0001
Words	47%	24%
Numbers	37%	56%
No preference	17%	20%
Agree uncomfortable with health information that has a lot of numbers or statistics	74%	43%	.0008
Agree depend on numbers and statistics to make decisions about health	39%	57%	.0036