Journal of Clinical Psychology in Medical Settings, Vol. 1, No. 2, 1994

Readability of Prescription Drug Labels by Older and Younger Adults J a n a n AI-Awar S m i t h e r I and Curt C. B r a u n 2

This experiment investigated the readability of medication labels using various combinations of fonts, sizes, and weights of print. Older and younger adult reading speeds and error rates for different print types were compared for both bottle and flat labels. Findings of the study indicated significant differences in readbTg speed, error rates, and subjective evaluations for print characteristics investigated. Furthermore, the effects of these print characteristics varied for different age groups. Implications of these findings for the design of medication labels for older adults are discussed. KEY WORDS: geriatrics; health psychology; pharmacy; compliance, reading legibility.

INTRODUCTION The elderly are the largest consumers of medication, expending 25% of all written prescriptions (Moore, Kalu, & Yavaprabbas, 1983). By the year 2030, this expenditure is expected to rise to about 40%. At the same time, medication noncompliance among the elderly has become a major health care concern (Morrow, Leirer, & Sheikh, 1988). Many researchers have investigated elderly noncompliance, with most of these studies focusing on issues such as the magnitude of noncompliance and the effects of instruction and labelling techniques (Leirer, Morrow, Pariante, & Sheik, 1988; Morrell, Park, & P o o n , 1990; Park, Morrell, Frieske, & Kincaid, 1992). Although some research addresses the readability of the medication label as a potential source of noncompliance, only one study (Zuccollo & Liddell, 1985) evaluated the legibility of several typeIThe University of Central Florida, Orlando, Florida 32816. 2The University of Idaho, Moscow, Idaho 83843. 149

1068.9583D4/0600-0149507.00/0© 1994PlenumPublishingCorporation

150

Smither and B r a u n

faces. Findings from that study indicated that 60% of the elderly participants had difficulty reading the labels produced at a hospital pharmacy. The lack of empirical research in the area of label legibility is surprising since a patient's ability to read the information on a label is, arguably, one of the first steps necessary for medication compliance. Research with Younger Subjects Although few legibility studies have included older adults, researchers have addressed legibility with children and young adults. Findings by Paterson and Tinker (1940) and Tinker (1963) suggest that, for young adults, the minimum legible type size is 8 points, with little improvement in reading speed expected above 11 points. Vanderplas and Vanderplas (1980) found that performance decrements related to typeface selection occurred only when unusual styles such as Old English were used. Other print factors that could affect legibility include font weight (bold vs. nonbold), presence or absence of serifs (i.e., small curves at the ends of the strokes), and character width (proportional vs. nonproportional). A study by Krulee and Novy (1986) showed bold type to be superior to nonbold for reading individual words. Tinker (1963), however, found no difference between bold and roman (nonbold) types in terms of reading speed. Tinker (1963) also found that although his readers expressed a strong preference for serifed type, there were very few performance differences. Payne (1967) and Beldie, Pastoor, and Schwarz (1983) have demonstrated the superiority of proportional over nonproportional spacing with both printed and digital text. Generalizing to Older Subjects Findings of legibility studies with older adults do not replicate those with younger ones. Poulton (1959, 1969) indicated that performance of older adults was affected negatively by small print and low illumination. Vanderplas and Vanderplas (1980) found a greater reading speed for Century Schoolbook font than for Helvetica, which is considered the standard for warning labels (FMC Corporation, 1985). These researchers also recommended the use of the 12- to 14-point size range, as opposed to the 8to 12-point size range that is the typical choice for younger persons. They concluded that because of the interaction of style-size combinations, line width, and leading (spacing between lines), the choice of print for any particular material requires task specific investigation.

Prescription Label Readability

151

Despite these differences in research results, there are some published guidelines for legibility for geriatric individuals based on data collected with younger individuals (Ralph, 1982). Considering the well-documented, agerelated declines in visual abilities among older adults, however, it is unclear that these recommendations would be applicable for developing print guidelines for older adults. Furthermore, evidence suggests that task specific analyses are necessary to ensure the proper selection of print characteristics (Vanderplas & Vanderplas, 1980). The present research was designed to assess the effects of age, font type, size, and weight on the readability of prescription medication labels. It was hypothesized that the print characteristics investigated will affect reading performance and preference and that these effects will differ for individuals of different age groups.

EXPERIMENT 1 Method

Subjects. The subjects were 19 younger adults [9 males (mean age = 22.12, SD = 1.93) and 10 females (mean age = 28.5, SD = 6.93)] and 20 older adults, over the age of 65 [10 males (mean age = 69.8, SD = 5.94) and 10 females (mean age = 72.3, SD = 5.53)]. The older volunteers were recruited from a learning institute for elders within a university setting. Younger volunteers were recruited from undergraduate psychology classes. All subjects were prescreened for reading ability, cognitive functioning, and near visual acuity. To participate, subjects had to be reading at a minimum of an eighth-grade level, their scores on the Digit Symbol test of the WAIS-R had to fall within the normal range for their age group, and those subjects with low vision were required to wear their corrective lenses during the experiment. Design. The study was a 2 (age) x 2 (gender) x 3 (font) x 3 (size) x 2 (weight) mixed-model design. The between-subject factors were age (young, old) and gender (male, female). The within-subject factors were font (Courier, Century Schoolbook, Helvetica), size (9 point, 12 point, 14 point), and weight (Bold, Roman). Material. The materials used were 72 standard size medication bottles (20 dram), 72 hospital pharmacy labels, a demographic questionnaire, a reading abilities test, the Digit Symbol test from the WAIS-R, and a Titmus Vision Tester. The selection of print types used was based on a review of the literature, as well as a survey of fonts used in major Central Florida pharmacies (Courier, 9 point, Roman). Courier and Century Schoolbook are serif fonts,

152

S m i t h e r and Braun

while Helvetica is a sans serif font. Century Schoolbook and Helvetica are proportional fonts, whereas Courier is a nonproportional font. The factorial combinations of font, size, and weight resulted in 18 label conditions. These conditions were then reproduced using four different sets of label information (LI). T h e LI sets were p r o d u c e d by the random assignment of the following pieces of information four times: (1) name of a physician, (2) name of a medication, (3) instructions, and (4) n u m b e r of refills. The end product of this process was four sets of 18 labeled medication bottles (see Table I). All labels were equated for content of information. Procedure. Subjects were first asked to complete the demographic questionnaire, a reading abilities test, and the Digit Symbol test. Next, each subject's near visual acuity was assessed. Each subject was then instructed to read the information on 18 labeled prescription bottles, one at a time, as quickly and accurately as possible. The order of presentation of the 18 bottles was randomized for each subject to reduce order effects. Before starting the experimental portion of the study, subjects were allowed to practice the procedure with three medication bottles. Subsequently, subjects were instructed to pick up a medication bottle and read the label information aloud. Time to read was measured in 0:01 of a second using a stopwatch. Subject accuracy was also recorded. After reading all the labels, subjects were instructed to look at each label again and to judge the print on a 7-point scale according to (1) ease of reading (very easy-very difficult), (2) clearness of type (very sharp-not at all sharp), and (3) appropriateness of size (too small-too large). Table 1. Examples of Font Type, Weight, and Size Point size Font

9

12

14

Bold Courier Century Schoolbook Helvetica

Refill

Refill

Refill

Refill

Refill"

Refill Refill

Refill

Refill Roman

Courier Century Schoolbook Helvetica

Refill b

Refill

Refill

Refill

Refill

Refill

Refill

Refill Refill

aRecommended font, size, and weight combination for use on bottles and flat labels. bprint type used by major Central Florida pharmacies.

Prescription Label Readability

153

Results

Reading Speed. A 2 x 2 x 3 x 3 x 2 mixed-model analysis of variance was performed on the reading speed data and indicated significant main effects for four independent variables: age, font, weight, and size (see Table II). Older adults were significantly slower than younger adults, and Roman weight was significantly slower to read than Bold. Range tests of the main effect of font indicated that Courier required significantly longer time to read than Helvetica and Century Schoolbook (p's < .0001). Similar range tests for size indicated that 9-point print required significantly more time to read than 12- and 14-point print (p's < .00001). The analysis also revealed several significant interactions. However, only those interactions that include age effects are included in this report (see Table II). Tests of simple effects for the weight × age interaction indicated no effects of weight for older adults. The three-way interaction of font x size x age was broken down by age group. Tests of simple effects performed on the separate interactions indicated that older adults performed best with the 12-point Century Schoolbook print. This interaction also indicated that 9-point was particularly slow with the Courier font, especially for the older subjects. Performance Errors. A similar A N O V A of subject error data indicated significant main effects for the three independent variables of age, font, and size (see Table II). Older adults committed more errors than their younger counterparts. Range tests indicated that more errors were committed with the Courier font than with Century Schoolbook or Helvetica (p's < .01) and that more errors were committed with the 9-point size than with the 12- or 14-point size (p's < .05). The analysis also revealed significant age interactions of font × age, weight x age, and font × weight x age (see Table II). Tests of simple effects for the font × age interaction indicated that older adults committed more errors with the Courier font than with Century Schoolbook or Helvetica (p's < .01) and that older adults committed more errors than younger adults with the Courier font (p's < .001). Furthermore, the weight x age interaction showed that older adults committed more errors with the Bold than the Roman print (p's < .01) and more errors with the Bold print than younger adults (p's < .001). The nature of this interaction indicated no effects of weight for younger adults. The three-way interaction was broken down by age group to form two font × weight interactions. Tests of simple effects of these interactions indicated that older adults committed more errors with the Courier Bold print than with the Courier Roman (p's < .001) and that older adults, when

2,68 1,34 2,68 1,34 2,68 2,70 2,70 2,68 2,68

Font (F) Weight (W) Size (S) W x A W x S x A Size (S) F x W x A

Font (F) Size (S)

Reading ease

Size

*p <.05.

Clearness

**p <.01.

1,35 2,70 1,35 2,70 1,35 4,140 1,35 2,70 2,70 2,70 1,35 2,70

Age (A) Font (F) Weight (W) Size (S) W x A F x Sx A Age (A) Font (F) Size (S) F ×A W ×A F x W x A

Reading speed

Errors

df

Experiment 1

Source

Variable

Variable

10.97"* 153.97"*

8.30** 15.45"* 12.84"* 5.92* 3.%* 80.13"* 5.04*

df

1,30

Weight (W)

Errors

Font (F) Weight (W) Size (S) FxWxA Font (S) Size (S) WxSxA

Clearness

Size

Font (F) Size (S)

Reading ease

3.83* 38.31"*

8.49** 4.51" 31.09"* 4.18" 3.26* 225.79** 4.90*

2,60 1,30 2,60 2,60 2,60 2,60 2,60

6.97*

20.91"* 8.17"* 8.94** 9.12"* 5.36**

F

2,60 2,60

1,30 2,60 2,60

Age (A) Font (F) Weight (W) Size (S) F x A 1,30 2,60

Experiment 2 Source

Reading speed

Subjective data

7.91"* 16.24"* 4.83* 15.39"* 4.33* 3.81"* 5.95* 5.68** 4.09* 4.68* 7.69** 3.13"

Reading speed and errors

F

Table II. F Ratios for Analysis of Experiment 1 and 2 Data

-I

g

Prescription Label Readability

155

reading Bold print, committed more errors with the Courier font than either Century Schoolbook or Helvetica (p's < .001). Subjective Ratings. Separate ANOVAs were performed for the three subjective scales of ease of reading, clearness of type, and appropriateness of size. Reading Ease. With respect to ease of reading, the analysis indicated significant main effects for font, weight, and size (see Table II). NewmanKeuls range tests indicated that subjects perceived Century Schoolbook (p's < .01), Roman weight, and 14-point print (p's < .0001) as the easiest to read. A test of simple effects of the weight x age interaction indicated that older subjects perceived Bold print to be more difficult to read than Roman (p's < .0001). Older subjects also felt that Bold print was more difficult to read than younger subjects did (p's < .05). The weight × age × size interaction was broken down by age group. Tests of simple effects indicated that, for older and younger adults, 14-point Bold was perceived as easier to read than 12-point Bold, which was, in turn, perceived to be easier to read than 9-point Bold (p's < .0001). With the Roman print, however, older and younger adults felt that the 9-point print was more difficult to read than either the 12- or the 14-point print (p's < .0001). Older adults also found both the 9- and the 12-point prints to be easier to read in Roman than in Bold type. Clearness. The analysis of ratings of clearness indicated a significant main effect of size and a three-way interaction of font × weight x age (see Table II). Range tests for size indicated that subjects judged the 14-point size to be sharper than the 12-point size, which was also judged to be sharper than the 9-point size (p's < .01). Tests of simple effects of the three-way interaction, again broken down by age group, indicated that older adults felt that Helvetica Roman was sharper than Helvetica Bold (p's < .05) and that Century Schoolbook Bold was sharper than Courier Bold (p's < .05). Size Appropriateness. The analysis of the appropriateness of size ratings indicated main effects for font and size (see Table II). Range tests of these main effects indicated that subjects felt that the size of the Century Schoolbook font was more appropriate than Helvetica, which in turn was considered more appropriate than Courier (p < .05). Subjects also felt that the 14-point print was more appropriate than the 12-point print, which was also considered more appropriate than the 9-point print (p < .001). The curvature of a typical medication bottle requires the reader to manipulate the label and also affects character and line width, surface glare, and contrast. Consequently, the results of the above study may not be generalizable to labels placed on a flat container. Therefore, a second experiment was designed to assess any differences in legibility that relate to the curvature of the label surface.

156

Smither and Braun

EXPERIMENT 2 Method

In Experiment 2, label information was manipulated in the same way as it was in experiment one; however, each prescription label was placed on a flat cardboard surface. Subjects. The subjects were 18 younger adults [8 males (mean age = 26.13, SD = 8.15) and 10 females (mean age = 21.7, SD = 0.949] and 16 older adults, over the age of 65 [8 males (mean age = 71.25, SD = 5,80) and 8 females (mean age = 70.6, SD = 9.56)]. These subjects were different from those in experiment 1; however, they were recruited from the same subject pools. As was the case with the first experiment, all subjects were screened for reading ability, cognitive functioning, and near visual acuity prior to participation. Results

Reading Speed. A 2 x 2 x 3 x 3 x 2 mixed-model A N O V A of the reading speed data indicated significant main effects for the four independent variables of age, font, weight, and size (see Table II). Older adults were slower than younger adults, and Roman weight was slower to read than Bold. Newman-Keuls range tests of the main effects of font and size indicated that Courier font was slower to read than Helvetica and Century Schoolbook (p's < .05) and that 9-point print was slower to read than 12or 14-point print (p's < .01). The analysis also revealed a font x age interaction (see Table II). A test of simple effects indicated that, for older adults, Courier was slower to read than both Century Schoolbook and Helvetica (p's < .005). Performance Errors. An A N O V A performed on subject error data indicated a significant main effect for weight IF(l, 30) = 6.97, p < .05], with subjects committing more errors with Roman than with Bold print. Subjective Ratings. Separate ANOVAs were performed for the three subjective scales of ease of reading, clearness of type, and appropriateness of size. Reading Ease. The ease of reading analysis indicated two significant main effects of font and size (see Table II). Range tests for the main effects of font and size indicated that subjects felt that Century Schoolbook and Helvetica fonts were easier to read than Courier (p's < .05) and that the 12- and 14-point prints were easier to read than the 9-point print (p's <

.0001).

Prescription Label Readability

157

Clearness. The clearness of type analysis indicated significant main effects of font, weight, and size (see Table II). Subsequent range tests revealed that subjects perceived Courier and Helvetica to be sharper than Century Schoolbook (p's < .003). Subjects also reported that 14-point print was sharper than 12-point print, which was also sharper than 9-point print (p's < .01). Subjects also judged the Bold print to be sharper than the Roman print. Furthermore, the analysis indicated a significant font x weight × age interaction (see Table II). This interaction was broken down by age group. Tests of simple effects revealed that older adults perceived Helvetica Bold as sharper than Helvetica Roman (p's < .005) and that, with Bold print, Helvetica and Courier were sharper than Century Schoolbook (p's < .01). Appropriateness of Size. The analysis of the appropriateness of print size ratings revealed main effects for font and size (see Table II). Subsequent range tests indicated that subjects judged the size of the Century Schoolbook font to be more appropriate than that of Helvetica (p's < .05) and the 14point print to be more appropriate than the 12-point print, which was also judged to be more appropriate than the 9-point print (p's < .0001). The analysis also revealed a significant weight x size x age interaction (see Table II). This interaction was broken down by age group. Tests of simple effects indicated that older adults felt that the size of the Bold 12-point was more appropriate than that of the Roman 12-point and that the size of the Roman 14-point was more appropriate than that of the Bold 14-point (p's < .0001).

DISCUSSION The results of the first experiment suggest that Century Schoolbook 12-point Bold or Roman print is the most appropriate choice for older readers of prescription bottle labels. Analysis of the reading speed data for the fiat labels revealed findings somewhat similar to those for the bottle labels (see Table II). The effects of error and subjective judgments with the flat presentations, however, appear to be fewer than arid different from those with the bottle presentations. These differences suggest that choice of print may be affected by the expected mode of presentation. In the case of fiat labels, the results indicate that Century Schoolbook or Helvetica 12- or 14point Bold print would be an appropriate choice. For a print type that could be applied on either bottle or box, the compromise would be a Century Schoolbook 12-point Bold print. Results of both experiments also suggest that the Courier 9-point print used by the four main pharmacies in the area where the study took place--Central Florida--appears to be the most difficult to read by all participants and is especially difficult to read for the older age group.

158

Smither and Braun

The overall findings of our investigations support previous readability research. As expected, the size of print affected performance. Moreover, although weight did affect reading speed in both the bottle and the flat conditions, the weight x age interaction in the bottle condition indicated that this effect applied only to the younger readers. In addition, a main effect of weight for errors appeared only in the fiat label condition. These findings provide support for the conclusion that the choice of print for any particular material requires task specific investigation (Vanderplas & Vanderplas, 1980) and, also, for the idea that it is inappropriate to generalize some findings from one age group to another. Indirect findings from this study relate to character width and the presence of serifs. As with previous research (Beldie et al., 1983; Payne, 1967), proportional spacing (Century Schoolbook and Helvetica) was found to be superior to nonproportional spacing (Courier). Furthermore, the effects of serifs on legibility were shown to be minimal in that Century Schoolbook (a serif print) and Helvetica (a sans serif print) yielded equivalent speed and accuracy performance data. In the same vein, Century Schoolbook and Courier, both serif prints, yielded significantly different performance data from each other. This research did not investigate readers' understanding or retention of the information presented on the different labels. We believe, however, that comprehension performance with low-readability print is likely to require more cognitive resources. Consequently, because of age-related declines in processing resources, any increased encoding demands caused by the lowreadability print may have more serious implications for older readers than for younger ones. Future research into the comprehension of medication information should therefore take the readability of the label into account. Results of this research are immediately applicable to medical settings, in which psychologists can communicate these findings to their colleagues. Readability of medication labels can be discussed as a quality assurance issue, and psychologists can provide print guidelines to their medical center pharmacies. In addition, clinical psychologists in collaboration with geriatricians and pharmacists can initiate research projects that investigate the relationships among readability, comprehension, and compliance to treatment regimes by elderly prescription medication consumers.

REFERENCES Beldie, I. P., Pastoor, S., & Schwarz, E. (1983). Fixed versus variable letter width for televised text. Human Factors, 25, 273-277. FMC Corporation (1985). Product safety sign and label system. Santa Clara, CA: FMC Corp.

Prescription Label Readability

159

Krulee, G. K., & Novy, F. (1986). Word processing and effects of variability in type fonts. Perceptual and Motor Skills, 62, 999-1010. Leirer, V. O., Morrow, D. G., Pariante, G. M., & Sheikh, J. I. (1988). Elders' nonadherence, its assessment, and computer assisted instruction for medication recall training. Journal of the American Geriatrics Society, 36, 877-884. Moore, S. R., Kalu, M., & Yavaprabbas, S. (1983). Receipt of prescription drug information by the elderly. Drug Intelligence and Clinical Pharmacy, 17, 920-923. Morrell, R. W., Park, D. C., &Poon, L. W. (1990). Effects of labelling techniques on memory and comprehension of prescription information in young and older adults. Journal of Gerontology, 45, 166-172. Morrow, D. G., Leirer, V. O., & Sheikh, J. I. (1988). Adherence and medication instructions: Review and recommendations. Journal of the American Geriatrics Society, 36, 1147-1160. Park, D. C., Morrell, R. W., Frieske, D., & Kinkaid, D. (1992). Medication adherence behaviors in older adults: Effects of external cognitive supports. Psychology and Aging, 7(2), 252-256. Paterson, D. G., & Tinker, M. A. (1940). How to make type readable. New York: Harper. Payne, D. E. (1967). Readability of typewritten material: Proportional versus standard spacing. Journal of T~pographic Research, 1(2). Poulton, E. C. (1959). Effects of printing types and formats on the comprehension of scientific journals. Nature, 184, 1824-1825. Poulton, E. C. (1969). Skimming lists of food ingredients printed in different sizes. Journal of Applied Psychology, 53, 55-58. Ralph, J. B. (1982). A geriatric visual concern: The need for publishing guidelines. Journal of the American Optometric Association, 53(1), 43-50. Tinker, M. A. (1963). Legibility of print. Ames: Iowa State University Press. Vanderplas, J. M., & Vanderplas, J. H. (1980). Some factors affecting legibility of printed materials for older adults. Perceptual and Motor Skills, 50, 923-932. Zuccollo, G., & Liddell, H. (1985). The elderly and the medication label: Doing it better. Age and Aging, 14, 371-376.