Elias J. Duryea, Ph.D.
Department of Health, Exercise & Sports Sciences, University of New Mexico
A paper presented at the Annual Meeting of the Society for the Scientific Study of Sexuality (SSSS - WR), Kona Kai Resort, Shelter Island San Diego, April 10-13, 2008. This research was supported by Sabattical Leave from the University of New Mexico to the University of Hawaii-Manoa & the University of California, 2007.
There is little debate in the behavioral science and medical literature that consumers of health-related information are routinely confused by conflicting, complex, unclear and/or inconsistent research findings ( Institute of Medicine, March 2004: Health Literacy: A Prescription to End Confusion, Washington, D.C. National Academies Press). While little research has been done in this area related to sexuality research findings, some new research focusing on “health literacy” and “health risk communications” appears to have direct relevance for sexuality investigators (Wolf et al., The Emerging Field of Health Literacy Research, Am J Health Behav, 2007, v31, Supplement 1). For example, much of empirical sexuality research is quantitative in nature and presents results with language such as “probability of”, “odds for” and “increased risk for”. Yet we have little data on how well the average consumer of sexual information actually understands the concept of probability or odds or how investigators actually calculate or derive “increased risk”. Weare now at a critical juncture in sexual science which allows researchers to stop and assess how proficient our audiences are at comprehending research reports and how best to make our outcomes more clear and thus compelling. This emerging research field has been referred to as “quantitative health literacy” or when referring to quantitative sexual information, as simply, sexual numeracy. For clarity this paper uses young people ages 14 to 18 as the reference point, although as the sexual numeracy field develops these propositions may well apply to different age cohorts.
Sexual numeracy can best be defined or operationalized as the level of understanding or comprehension an individual possesses when presented with sexuality information that is mostly quantitative in nature. Since it is a new conceptual area there is little if any empirical documentation of its psychometric properties although Duryea et al., have begun developing a valid and reliable instrument to measure this skill domain (Duryea et al., 2008, Paper to be presented at the American Public Health Association Annual Meeting, San Diego, October, 2008).
While illustrations of different sexual numeracy scenarios will be examined in subsequent sections, a brief sampling of sexual numeracy areas could include the following:
In each of these scenarios the individual has received quantitative information related to sexual behavior and its consequences. While young people graduating from high school are supposed to have studied and shown competency in understanding odds, probabilities and related quantitative concepts in math courses, we do not know how well most actually DO comprehend them in real-life situations. This unknown level of skill in sexual numeracy among our youth should be of concern to all sexuality educators as well as researchers because lack of such proficiency has an array of both social as well as health consequences (Schwartz et al,Role of numeracy in understanding the benefits of mammography,Ann Intern Med, 1997, 127(11), 966-972).
There is a burgeoning wealth of research emerging examining how well adults understand health and numerically presented health information. The entire issue of the American Journal of HealthBehavior was recently devoted to the state-of-the-art research in this domain (2007, 31(Suppl 1). Unfortunately, none of the authors presented data specifically addressing sexual-based literacy or sexual numeracy.
One startling finding reported, however, was the finding from the National Adult Literacy Survey that almost a quarter of US adults were below basic proficiency in “quantitative literacy skills” (QLS). Only 33% were rated as having basic skills (Kirsch et al, US Dept of Education, 1993, NCES 93275). Even more alarming was a finding in a study of “highly educated” samples that only 55% of respondents correctly answered how many times a die would come up even in 1000 rolls (Lipkus et al, Med Dec Making, 2001,21(1), 37-44).
Medical care investigators in particular have conducted many studies on numeracy in various kinds of patients- and its ramifications. While this literature is beyond the scope or purview of this paper, a common theme is that patients with low numeracy skill have a range of problems weighing the cost to benefit ratio of various procedures. In one study, cancer patients who answered a numeracy question incorrectly over-estimated the benefits of ineffective treatment options for their cancer. Those who answered the numeracy question correctly did not make this mistake (Weinfurt et al., Correlation between patient traits & expectations of benefit from phase I trials, 2003, Lancet, 359, 1124-1125). Clearly such deficiencies have important implications for how one chooses health and specifically medical decision options. If people are not adept at accurately understanding and thus interpreting quantitatively the risks versus utilities of various treatments before them, the chances increase that they may make a choice which is not in their best medical interest (Dewalt et al, Literacy & Its Relation to Self-Efficacy..& Participation in Medical Decision Making, Am J Health Behav, 2007,31(S27).
Moreover sexuality researchers, in particular, have to confront the very real possibility that this lack of proficiency may have severe and life-altering consequences not only in the short-term but for young peoples future lives. These potential endpoints or consequences are examined briefly next.
Endpoints of Low Sexual Numeracy
The previously presented scenarios offer a sampler of how low sexual numeracy skill may impact young peoples lives. Every sexuality teacher/educator, for instance, knows the pedagogic challenge of trying to teach students how to comprehend “failure/effectiveness” rates of various contraceptives. While the new options in the domain of oral contraceptives remain extremely favorable with regard to averting an unplanned pregnancy, young people may feign comprehension of what these estimates mean and simply do “what everyone else is doing”. Because there is so much more to such a personal and far-reaching choice than simply the “failure rate”, young people need proficiency in exploring what the rate means numerically, how it is derived and how much weight it carries along with other considerations (e.g., medical status, preventing SDI’s).
Case of IPV -
Yet contraceptive selection is only one important decision making area where the ramification of having low sexual numeracy can be a potent and negative endpoint.
Consider the situation where a young couple slowly but steadily drift into a relationship that starts to produce intimate partner violence (IPV). The young woman may hear or read that the “incidence” of IPV in her age group is 2.5 events / month / couple. How should she view this quantity ? What does incidence mean compared to an older college friend of hers who tells her that the “prevalence” of IPV is actually double that ? Is 2.5 beyond “normal expectancy”, another term she recently read about in a sociology text.
How do supposed experts derive this normal expectancy? In an all to frequent and worse case scenario the young woman does not have the quantitative or academic foundation to understand or critique this information and decides not to interpret the episodes of IPV as health-compromising or “abnormal”. The endpoint here may be the routinization of an abusive and eventual life-threatening relationship.
Case of Genetic Screens –
Another area where young people, in fact all couples starting a family, need sound numeracy is in understanding probability or odds of a child having a genetic condition. If, for example, the genetic counselor informs a young couple that the probability of their baby having cystic fibrosis is 20%, what does that mean ? If the counselor said that the odds of having a CF baby were 1 in 5, would that estimate mean something different? Regardless, the couple may make a decision without asking the genetic counselor to fully explain what these measures mean and how they are derived. 1 Fortunately modern hospitals’ genetic counseling departments generally perform exceptionally in this area. The potential problem remains, however; that a young couple may not fully understand these estimates quantitatively and thus their eventually decision may not be fully informed.
Case of HIV/AIDs –
Without much argument sexuality specialists concur that young people newly diagnosed with sero-positivity for the HIV require precise and compelling advice on their immediate course of action. Consider that while still theoretically fatal, HIV if diagnosed early, in a healthy young person and quickly entering treatment with state-of-the-art anti-retroviral drugs, the fatality rate need not be a major concern. Of course, in too many poor and/or disenfranchised urban regions of the US (not to mention the world), young people do not have these advantages. In such terrain a young person may hear the word “fatal” or “terminal” or even “casefatality rate” associated with AIDs and simply conclude that s(he) has no option for survival.
How do young people interpret and react to the supposed absolute finality of such concepts? Do they pause to reflect and consider that a case fatality rate (CFR) has a counterpart (the survival rate) which offers some hope in the form of a numeric estimate of those who do not die ? Even if they do ask questions about the CFR and are told that it is 30%, what are they to make of it ? Is it high and thus risky, low or somewhere in between ?
Nowhere was this endpoint or consequence more evident than in the spike in unprotected sex among San Francisco’s gay men following reports that the fatality rate for gay men in the San Francisco Gay Men’s Health Study had increased despite previously high rates of “protected sex” among these men (Winkelstein et al., San Francisco Gay Men’s Health Study, Gateway.nlm.nih/gov.). Asked why they had stopped protecting themselves many simply stated that “it doesn’t matter if you protect, its still fatal in the end”. In fact, the rise in deaths was related to the long duration and advanced stage of the disease in many of these men - not because they were simply destined to die regardless of any behavioral protection they enacted.
While there are numerous other related examples of low sexual numeracy influencing sexual health behavior, a constant factor remains: sexuality specialists need to tailor their sexual information, especially when quantitative, to the numeracy level of recipients. How to construct and display quantitative sexuality information in ways that assist individuals in better understanding this type of information is the focus of the next section.
The discipline of communication has long sought to design effective channels that optimize the comprehension of presented material. In relatively recent communications literature the field of health risk communication has emerged. In this field researchers test out what types/combination of format, order, portrayals and written prose best suits the traits of the target audience. Unfortunately, little of this research base has explored how these variables may be used to assist young people gain greater understanding of sexuality information that is numeric in nature. Still researchers in health literacy and health risk communication have laid a foundation for such investigations (A Fagerlin et al., Am J Health Behavior, 2007, Suppl 1,s47-56: Making Numbers Matter: Present Research in Risk Communication). A summary review of how their findings may relate to sexuality-based quantitative information follows. Frequencies or Percentages?
Researchers suggest that where the audience is regarded as low in numeracy skill, frequencies should be used to convey the information over percentages. Where the audience is said to be high numeracy then the converse should be followed (Schapira et al., Med Dec Making, 2001,21,459-467, Freq or Probability ?). If the numeracy level of the target audience is not known then researchers suggest presenting both with frequencies first. Seasoned investigators in behavioral science often report findings in this last format by showing in columns the number that occurred (frequency count) and then in ( ) , showing what % it is of the total, i.e., 66 (92). Yet for young audiences and when the reference is sexual data we have little information on what works most effectively. A conservative approach here is to show data in both forms with a clear, brief narrative explaining what the numbers mean.
Bar Graph v. Line Figure v. 2 or 3 – Dimensional Histograms?
Research also indicates that individuals with low numeracy grasp information presented in basic bar graphs (1 - dimensional) better that other formats. Data shown in line figures with typical x and y ordinates and information in 2 or 3 – dimensional histogram formats are better suited for those with high numeracy skill. Curiously, researchers have found that while numeracy may be related to understanding it may not always be associated with an individuals preference (Zacks et al., J Exp Psychol Appl, 1998,4:119-138, Reading Bar Graphs: Effects of Extraneous Depth Cues & Graphical Context.).
It should be cautioned that while health risk communications and health literacy researchers have developed measures to estimate health risk, health literacy and health numeracy ability, the same is not yet available for sexuality-based information, especially with 14 to 18 years olds. Investigators at the University of New Mexico have one of the first efforts in this area (EJ Duryea, Presentation to the American Public Health Association Conference3, San Diego, October 2008). The content and psychometric tests of this instrument is in development using first-year university students – a prime audience for sexually-based health risk.
Choosing Time Spans: Longer or Shorter?
Investigators have found that the best way to present “risk” information on health choices (sexual should not differ), involves selecting the appropriate time span to show the risk. For example, informing people they have a 33% risk of severe injury if they do not use seat belts in a lifetime of driving yields greater usage than telling them the much smaller risk for a single trip (Schwalm ND et al, Perceptronics, Woodland Hills,CA.,1982, Report No.-PFTR-1100-82-83.) When women were asked how they wished to view their risk of breast cancer, long span or short, they selected longer span: women of all ages converged on wanting that information in 10 year time spans as opposed to annual risk estimates. Their rationale in post-study follow-up: a decade was a span of time in which one normally makes major life plans (prior Schapira, 2001).
Present “Gain” or “Loss” Frames?
The “framing effect” refers to how quantitative information is presented to readers. For example, when only told that a treatment/procedure has a 90% chance of being effective (gain frame), a majority of patients choose it. Yet if they are told that it has a 10% chance of it not being effective (loss frame), a majority decline it (McNeil et al., N Engl J Med, 1982,306:1259-1262: On the elicitation of preferences for alternative therapies).
Researchers have found that low numeracy individuals prefer quantified ratings in a gain frame more than did high numeracy subjects. High numeracy subjects preferred either frame supposedly because of their ability to translate both frames equally (i.e., they quickly comprehend that the ratings regardless of how ‘framed’, are the same result (Peters et al, Psychol Sci, 2006, 17, 407-411: Numeracy & decision making).
The implications of how sexually – based numeric material is presented via these 2 frames are far reaching. Firstly, we have little information on how young people with low numeracy skills cognitively interpret quantitative information across these frames. One example being explored is to present young people with both low and then high numeracy proficiency the following scenario involving assessment of “condom effectiveness”:
You read that condom type X has a failure rate of 5% , while condom type Z has a success rate of 95 %, both are identical in all other respects (cost, appearance, feeling) which would you choose? X __ Z __
Secondly, researchers and especially sexuality educators and counselors, need to know not only what numeracy level students/clients are at but also how best to ‘frame’ numeric estimates so that individuals can make the best choice for themselves. Finally, framing research is not clear on how numeracy level relates to actual individual “preferences”: a low numeracy person may well choose a certain treatment not because of frame but because of some innate preference not even associated with how the option was framed.
Qualitative follow-up studies need to ask these individuals as well as high numeracy youth why and how they made their choices. Armed with such data sex educators, therapists and counselors will at minimum have some foundation for how to construct and present to students/clients sexual information that is quantitative in context.
To individuals without a background in behavioral science, epidemiology or public health, hearing that use of cigarettes with oral contraceptives over a long time period increases the risk for later health problems, seems commonsensical. Yet few understand how the quantification of “risk” is carried out. Young people often hear from health / sexuality teachers, adult authorities and others that if they do X it will double their risk of getting Y. Alcohol and motor vehicle crashes, unprotected sex and SDI’s and a plethora of related social and health admonitions in this area are routinely encountered by youth. Most probably do not even stop to reflect on what “increased risk” specifically implies.
We argue that lack of skill in this domain is related to young people’s tendency to say things like “I live for the moment”, “I’m not gonna stress on it” and thus act/feel that they are essentially “invulnerable”. If this hypothesis is true and an array of psychological and social theories posit that it is a major predictive factor in health-compromising actions (e.g., Health Belief Model, Protection Motivation Theory), then it makes sense for researchers and educators to invest some level of effort in explaining the concept of “increased risk”.
What “Increased Risk” Means-
The field of epidemiology has developed a large number of quantified expressions for how delineating “increased risk” (attributable, absolute, relative). Consider the following for what is probably the most common risk estimate: the relative risk ratio (RR). In its most basic form, for example, this simply means that the rate of the health problem (i.e.,unwanted pregnancy (UP)) in a “high risk group” (those who do not use condoms) is divided by the rate of UP in those who do use condoms (low risk group): 27 per 1000 population / 9 per 1000 population = 3 per 1000 population or RR = 3.0 times > risk
Those who fail to use condoms are thus said to have increased their risk for UP 3-fold relative to those who do use condoms. Much of preventive health and sexuality education rests upon this core concept: if one engages in health-jeopardizing activities, one realizes higher rates of associated health/sexual problems or “increased risk”.
The barrage of media reports that Americans regularly consume or are exposed to, refer to “risk” in an almost colloquial way: “new mothers with low levels of prenatal care have an increased risk for birth problems”, “use of sex supplements increases risk for health problems”. No less informative are government and academic reports: “use of OC’s linked to a lower risk of ovarian cancer” (Nation’s Health, 3/2008), “an estimated 33-54% of US women are at risk for domestic violence during their lifetime” (Am J Pub Health, 2008, 98(3). It is not surprising that young people may not comprehend this language, the general population as a whole may not fair much better. Some investigators contend that the dilemma of enhancing numeracy is even more critical in under-represented or disenfranchised populations (A Kalet et al., J Gen Int Med, 1994, 9(7): 402: How do physicians talk with their patients about cancer?)
In “hard-to-reach” or “hidden populations” the barriers to getting individuals numeracy skills increased are even more daunting. Not only are such populations lacking in general literacy but they are invariably poor, minority and or support-challenged. Despite these barriers some researchers have been able to infuse very creative and beneficial numeracy interventions. In Poipet town on the Thai/Cambodia border, for instance, sex workers are extremely vulnerable to HIV/AIDS. The CFDS a small Cambodian NGO recently intervened with brothel owners to help these women learn an array of life skills one of which was numeric “saving schemes”. In this specific skill-building component women learn first hand how to save an increment of their money to use for future educational pursuits. Prior to this training the women possessed virtually no numeric understanding of “income” and “savings”. Some of these sex workers have been able to actually leave the brothels and find self-sustaining jobs in other less-risky areas. The potential long-term impact on AIDS rates and personal medical cost for HIV is readily evident (Phay J. Int Conf AIDS, 2004, July 11-16, Abstract No. D11173).
In May of 2008 the Institute for Healthcare Advancement will hold yet another national Health Literacy Conference ( Irvine CA., May 1,2). It will not examine sexual numeracy or quantitative literacy but instead focuses on those who work in clinical settings with patients of low health literacy. The discipline of sexuality or sexual health is not part of the major conference program. Still it demonstrates how health literacy has risen to become a central and contemporary part of the national academic and medical arenas.
In San Francisco recently an innovative group of social activists designed and evaluated a “sexual health text messaging“service for San Francisco youth (Levine et al., SEXINFO, Am J Pub Health, 2008). It allows youth to text sexuality questions back and forth to a clearinghouse but does not address numeric queries.
Additionally, a recently published group randomized trial of HIV prevention among young males uses their fathers as the primary sexuality educator yet without coverage of sexual numeracy areas (DiLorio et al., REAL Men, Am J Pub Health, 2007, 97:1027).
It is readily clear that numerous creative programs and research trials are being developed and disseminated to young populations. What is strikingly absent unfortunately are components in these efforts to teach youth how to understand and correctly interpret sexual information that is quantitative in context.
Consider the following vignette:
…a young person is diagnosed with herpes at his/her school health services…s(he) recalls that herpes is considered incurable…(s)he is beset with excruciating stress over this fact…this weakens her ability to concentrate and function…in a subsequent meeting with a physician he is told of available treatment options… she hears “the probability of drug A causing negative side effects such as…is 40%”…”with drug B your risk for having complications such as …. is moderate with 40 out of 100, 000 in your age cohort experiencing it”…
Notwithstanding that many young people entering college are already deficient in quantitative skills, this person must now negotiate both stress AND a long-standing fear of anything statistical. Still treatment decisions must be made and the normal response here is probably for the patient to simply defer to the doctor. This almost innate trepidation of American youth (not to mention adults) when confronted with numeric information to defer to supposed authorities is at the root of patients not fully participating in their own health care decisions. Conversely, the young adult who is given practice in developing competence in understanding and evaluating this type of data may still be stressed but at minimum is not thoroughly daunted by the numerical information presented to her. He may even ask the M.D. for readings before choosing a treatment.
Young people today more than ever before possess exemplary technological skill: they seem to naturally have no fear of pursuing and mastering interfaced computer gaming, music uploading, internet searching, video and cell phone complexities and other technological advances. Yet if asked to calculate the % increase in college-aged chlamydia if the rate was 8/1000 students but is now 19 / 1000 students, most would likely be at a loss. Now is the time for research in this new sub area of health literacy to be more fully explored. With the burgeoning increase of interest in how we understand health information, sexual numeracy skill acquisition must also be a salient part of our educational, sexual and health advocacy.
1. As Gordis (Epidemiology, 3rd Edition, Saunders 2004) has explained: odds are made up of 2 probabilities, P1 or the numerator (chance of an event occurring) & P2 its denominator (chance of it not occurring).