Health Literacy, Pedometer, and Self-Reported Walking Among Older Adults

Abstract

Objectives. We examined the association of health literacy with physical activity and physical activity guideline adherence in older adults.

Methods. We used cross-sectional data from a 2012 population-based study in Alberta, Canada, assessing health literacy, and deriving moderate-to-vigorous physical activity (MVPA) and metabolic equivalent of task (MET) minutes per week from the Godin Leisure-Time Exercise Questionnaire, and steps per day via a pedometer.

Results. Mean age of participants (n = 1296) was 66.4 (SD = 8.2) years, 57% were female, and 94% were White. Nine percent had inadequate health literacy, and 46% met guidelines for self-reported physical activity and 18% for steps per day. Participants with inadequate health literacy had nonsignificant adjusted decrements of 58 MVPA minutes and 218 MET minutes per week and were less likely to meet physical activity guidelines (MVPA: odds ratio = 0.63; 95% confidence interval [CI] = 0.41, 0.97; P = .037; MET: odds ratio = 0.65; 95% CI = 0.42, 1.01; P = .057) compared with their health-literate counterparts. Such differences were nonsignificant for steps per day.

Conclusions. Inadequate health literacy was associated with less likelihood of meeting MVPA guidelines based on self-reported physical activity, but not based on an objective measure of steps per day.

Physical inactivity has been recognized as a modifiable risk factor for cardiovascular disease and a widening variety of other chronic conditions, which are mostly common in the elderly population. Evidence from 3 large systematic reviews of the benefits of physical activity on health suggests that regular aerobic activity and short-term exercise programs are associated with functional independence; a positive effect on primary prevention of several chronic conditions including heart disease, hypertension, stroke, type 2 diabetes, some types of cancers, dementia, and Alzheimer’s disease; as well as a reduction in total mortality among adults.^1–3 Because of the rapidly increasing proportion of older adults worldwide, which is imminently expected to exceed that of children aged 5 years or younger for the first time in history,⁴ and recent reports indicating that approximately 31% of adults worldwide are physically inactive,⁵ physical inactivity represents an increasing global public health problem.

Previous research has identified several factors that are directly or indirectly linked to physical activity in adults, including age, gender, income, self-efficacy, motivation, and safe, accessible locations for walking and other activities known to improve health.^6–9 Furthermore, physiological, cognitive, and social changes that occur with aging are also important factors to consider when one is examining physical activity in older adults; however, very little attention has been paid to how other unique factors might influence health behaviors in older adult populations. For instance, health literacy, a relatively new concept in health research, has been the focus of recent literature.

Health literacy (HL), which is commonly defined as the ability to obtain, read, understand, and communicate about health-related information needed to make informed health decisions,¹⁰ has been recognized as a stronger predictor of a person’s health than conventional correlates such as age, income, education, employment status, and race.¹¹ Inadequate HL disproportionately affects older adults and other disadvantaged groups (e.g., people with low levels of income and education, and those for whom English is a second language), adversely affecting several aspects of their health care management and outcomes.^12–14 Several studies have suggested an important link between HL and several health-related outcomes among community-dwelling elders, including poorer medication knowledge,¹⁵ poorer physical and mental health,¹⁶ higher hospitalization rates, and increased cardiovascular and all-cause mortality.^17,18 A behavioral pathway has been suggested to explain some of these associations, whereby HL was found to be associated with several health-promoting behaviors including physical activity,¹⁹ and consumption of healthy foods^20,21 that are potentially linked to these outcomes. Despite these findings, the evidence is inconclusive because of a paucity of studies and potential methodological limitations in existing literature.^22,23

The primary objective of this study was to investigate whether inadequate HL is associated with self-reported moderate and vigorous physical activity (MVPA) in community-dwelling older adults. We hypothesized that those with lower HL would not meet current physical activity guidelines. Secondarily, because walking is the most commonly reported form of MVPA among older adult populations, we also elected to explore this relationship by using objectively assessed ambulatory activity (walking) with a pedometer. We hypothesized that those with lower HL would report fewer steps per day.

METHODS

We used cross-sectional data from the Alberta Older Adult Health Behavior (ALERT) study for these analyses. The details of this study have been previously reported.²⁴ Briefly, random digit dialing was used to recruit a sample of older adults in Alberta, Canada. Individuals were eligible to participate if they were aged at least 55 years, able to walk unassisted, able to complete a telephone survey in English, community-dwelling residents of Alberta at the time of the study (i.e., not institutionalized), and could be contacted by direct phone dialing. Participants completed a sociodemographic, health behavior, and outcome measures questionnaire by using computer-assisted telephone interviewing. A pedometer was mailed to participants after they completed the questionnaire. The response rate for the telephone interview was 18.5% of 7013 eligible participants approached, with 70.1% who refused to participate and 11.4% who were deemed ineligible to take part in the study. The majority of the recruited sample (83%) submitted self-measured walking with the pedometer. We conducted data collection between September and November 2012.

Measures

For sociodemographic information, we collected data on age, gender, education (less than high school, completed high school, more than high school), annual household income (< Can $40 000; Can $40 000–Can $80 000; > Can $80 000), employment status (employed or unemployed/retired), ethnicity (White, Aboriginal, other), current smoking status (smoker, nonsmoker), self-reported height and weight (used to calculate body mass index [BMI, defined as weight in kilograms divided by the square of height in meters]), and number of chronic conditions (e.g., chronic obstructive pulmonary disease, heart disease, hypertension, diabetes, mood disorders).

We measured HL by using 3 brief screening questions (difficulty in understanding written information, confidence in completing medical forms, needing help in reading medical information) that have been previously used and validated in older adults.^25–27 Each question was scored from 1 to 5 with higher scores indicating lower HL, and a cut-off point of 9.0 or higher for the weighted summative score (range = 3–15) indicating inadequate HL.^25,28

We used the Leisure Score Index of the Godin Leisure-Time Exercise Questionnaire to estimate physical activity.²⁹ The Leisure Score Index contains 3 questions assessing the average frequency and duration of light, moderate, and vigorous intensity physical activity that lasted at least 10 minutes during leisure time over a typical week during the past month. We computed the number of minutes per week for each activity level by multiplying the frequency and duration of activity in that level. We computed MVPA minutes by adding the weekly minutes of moderate physical activity and 2 times the weekly minutes of vigorous physical activity, as recommended by the 2008 American Public Health Guidelines³⁰ and others.³¹ We computed a total metabolic equivalent of task (MET) score by adding weekly minutes of moderate physical activity multiplied by 4.0 METs and weekly minutes of vigorous physical activity multiplied by 7.5 METs.³¹ One minute of vigorous physical activity is equivalent to 1.875 minutes of moderate activity (7.5/4.0) based on the average MET levels for vigorous activity (MET level = 7.5) and moderate activity (MET level = 4.0). This weighting provides more credit for participating in vigorous activity. Meeting guidelines for self-reported physical activity was determined by a score of 150 or more for the MVPA minutes per week, and 600 or more for the MET minutes per week.³¹

Objective walking behavior was assessed with a step pedometer (StepsCount SC-01, Deep River, Ontario, Canada). Participants were asked to start tracking their daily steps over a 3-day period the following morning after receiving the mailed package, and were encouraged to wear the pedometer on 1 weekend day. Acquiring between 7000 and 10 000 steps per day is considered an acceptable daily physical activity target and beneficial for overall health.³² To establish whether participants were meeting guidelines based on pedometer steps, we used a cut-off of 8000 or more steps per day of the average of the 3-day measurements.³³ We analyzed both self-reported physical activity and pedometer-measured walking as continuous and categorical outcomes.

Data Analysis

We calculated descriptive statistics for the overall sample and by HL level. We examined both self-reported physical activity by using the MVPA and MET minutes and pedometer-measured walking in the overall sample and by HL level. We tested differences in participants’ characteristics and physical activity levels by HL by using the χ² test or t test as appropriate. To examine the independent associations of HL with physical activity measurements, we used multivariable linear regression models (for continuous scores of self-reported physical activity and pedometer-measured walking) and logistic regression models (for binary outcome—i.e., meeting guidelines for self-reported physical activity and pedometer-measured walking) with HL as the main explanatory variable, and adjusted for age, gender, education, income, employment status, ethnicity, smoking status, BMI, and number of chronic conditions. Because the outcomes were not entirely normally distributed, we used generalized linear models with γ distribution as a sensitivity analysis.

We also calculated effect size (ES) of differences (ES = difference/overall SD) between HL groups for all continuous outcomes for a better understanding of the magnitude of these differences (ES: 0.2–0.49 = small; 0.5–0.79 = moderate; ≥ 0.8 = large³⁴). All participants provided steps data on all 3 days, and the percentage of missing data in the survey measures was less than 1%,²⁴ so we employed mean imputation for missing data in continuous variables. We completed data analysis with Stata version 13.0 (StataCorp, College Station, TX).

RESULTS

Mean age of participants (n = 1296) was 66.4 (SD = 8.2) years, 57% were female, 94% were White, slightly more than two thirds (67%) had more than a high-school education, one third had an annual household income of more than $80 000, and 59% were unemployed or retired (Table 1). The mean BMI of this population was 27.2 (SD = 5.1), with more than half considered overweight or obese (i.e., BMI > 25), and the average number of chronic conditions was 2.5 (SD = 1.9) with almost two thirds (64.4%) reporting 2 or more comorbid conditions. We found inadequate HL among 9% of the sample. Participants with inadequate HL were older, less educated, with lower income, more likely to be unemployed, and reported more chronic conditions compared with those with adequate HL (Table 1).

TABLE 1—

General Characteristics of Participants, Overall and by Health Literacy Level: Alberta Older Adult Health Behavior Study; Alberta, Canada; 2012

Characteristics	Overall (n = 1296), No. (%) or Mean ±SD	Adequate HL (n = 1183; 91.3%), No. (%) or Mean ±SD	Inadequate HL (n = 113; 8.7%), No. (%) or Mean ±SD	P
Demographics
Age, y	66.4 ±8.2	66.2 ±8.0	68.7 ±9.3	.002
Age group, y				.016
54–64	640 (49.4)	596 (50.4)	44 (38.9)
65–75	435 (33.6)	395 (33.4)	40 (35.4)
> 75	221 (17.1)	192 (16.2)	29 (25.7)
Female gender	741 (57.2)	682 (57.7)	59 (52.2)	.26
Education				<.001
< high school	144 (11.1)	115 (9.7)	29 (25.7)
Completed high school	289 (22.3)	256 (21.6)	33 (29.2)
> high school	863 (66.6)	812 (68.7)	51 (45.1)
Income, Can$				.003
< 40 000	305 (23.5)	269 (22.7)	36 (31.9)
40 000–80 000	334 (25.8)	297 (25.1)	37 (32.7)
> 80 000	415 (32.0)	395 (33.4)	20 (17.7)
Prefers not to answer	242 (18.7)	222 (18.8)	20 (17.7)
Employment – unemployed or retired	769 (59.3)	691 (58.4)	78 (69.0)	.028
Race/ethnicity				.58
White	1221 (94.2)	1117 (94.4)	104 (92.0)
Aboriginal	16 (1.2)	14 (1.2)	2 (1.8)
Other	59 (4.6)	52 (4.4)	7 (6.2)
Lifestyle behaviors
Current smoking	160 (12.4)	146 (12.3)	14 (12.4)	.99
BMI, kg/m2	27.2 ±5.1	27.1 ±5.1	27.5 ±5.0	.411
BMI categories (kg/m2)				.43
Underweight (< 18.5)	16 (1.2)	15 (1.3)	1 (0.9)
Normal weight (18.5–24.9)	449 (34.7)	413 (34.9)	36 (31.9)
Overweight (25–29.9)	533 (41.1)	490 (41.4)	43 (38.1)
Obese (≥ 30)	298 (23.0)	265 (22.4)	33 (29.2)
Number of comorbidities	2.5 ±1.9	2.4 ±1.8	3.0 ±2.1	.002
Number of comorbidities – categories				.09
0	168 (13.0)	159 (13.4)	9 (8.0)
1	294 (22.7)	273 (23.1)	21 (18.6)
≥ 2	834 (64.4)	751 (63.5)	83 (73.5)

Note. BMI = body mass index; HL = health literacy.

Physical Activity in the Overall Sample and by Health Literacy

Overall, less than half of the participants met guidelines for self-reported physical activity based on suggested thresholds—46% for both MVPA and MET minutes (Table 2). However, based on pedometer-measured walking, only 18.6% were considered to meet guidelines for physical activity based on suggested thresholds. Participants with inadequate HL were significantly less likely to meet guidelines for self-reported physical activity, whereby only 31% of the inadequate HL group met these guidelines for both MVPA and MET minutes, compared with 48% and 47% for these measures, respectively, in the adequate HL group. Nonetheless, we did not observe these differences for pedometer-measured walking, whereby only 19% met guidelines for physical activity in both adequate and inadequate HL groups (Table 2).

TABLE 2—

Self-Reported and Pedometer Physical Activity Measurements, Overall and by Health Literacy Level: Alberta Older Adult Health Behavior Study; Alberta, Canada; 2012

	Physical Activity, Self-Reported			Physical Activity, Pedometer
Outcome	Overall (n = 1296), No. (%) or Mean ±SD	Adequate HL (n = 1183; 91.3%), No. (%) or Mean ±SD	Inadequate HL (n = 113; 8.7%), No. (%) or Mean ±SD	Overall (n = 1081), No. (%) or Mean ±SD	Adequate HL (n = 991; 91.7%), No. (%) or Mean ±SD	Inadequate HL (n = 90; 8.3%), No. (%) or Mean ±SD
MVPA
Total MVPA min/week	246 ±385	253.2 ±394	165 ±257
Meeting guidelines: ≥ 150 min/wk	600 (46)	564 (48)	36 (32)
MET min
Total MET min/wk	950 ±1470	979 ±1504	643 ±999
Meeting guidelines: ≥ 600 MET min/wk	590 (46)	554 (47)	36 (32)
Total steps/d				5163 ±3539	5215 ±3527	4827 ±3675
Meeting guidelines: ≥ 8000 steps/d				201 (19)	184 (19)	17 (19)

Notes. HL = health literacy; MET = metabolic equivalent of a task; MVPA = moderate to vigorous physical activity.

Health Literacy and Physical Activity Measurements

In unadjusted analysis, participants with inadequate HL had statistically significant and clinically important decrements of 88.3 in total MVPA minutes per week (P = .02; ES = 0.2) and 336 MET minutes per week (P = .02; ES = 0.2) compared with those with adequate HL (Table 3). We did not observe such differences for pedometer-measured walking, with a decrement of 389 steps per day (P = .32; ES = 0.1) for adults with inadequate HL compared with those with adequate HL skills. These results were consistent with those from γ regression analysis; MVPA minutes per week: exponentiated coefficient = 0.65; 95% confidence interval [CI] = 0.48, 0.88; P = .01; MET minutes per week: exponentiated coefficient = 0.66; 95% CI = 0.49, 0.88; P = .01; pedometer steps: exponentiated coefficient = 0.93; 95% CI = 0.80, 1.07; P = .30). Compared with adults with adequate HL skills, those with inadequate HL were much less likely to achieve guidelines for self-reported physical activity for both MVPA (odds ratio [OR] = 0.51; 95% CI = 0.34, 0.77) and MET minutes (OR = 0.53; 95% CI = 0.34, 0.80). Such an association was not observed for pedometer-measured walking (OR = 1.02; 95% CI = 0.59, 1.77).

TABLE 3—

Results of Linear and Logistic Regression Models of Physical Activity and Health Literacy: Alberta Older Adult Health Behavior Study, Alberta, Canada, 2012

	Unadjusted			Adjusteda
Outcome	b (95% CI)	OR (95% CI)	P	b (95% CI)	OR (95% CI)	P
Number of MVPA min/wk (n = 1296)	−88 (–163, –14)		.02	−58 (–132, 17)		.12
Number of MET min/wk (n = 1296)	−336 (–619, –52)		.02	−218 (–502, 66)		.13
Number of pedometer steps/d (n = 1081)	−389 (–1153, 376)		.32	39 (–668, 745.8)		.92
Meeting MVPA min/wk guidelines (n = 1296)		0.51 (0.34, 0.77)	.001		0.63 (0.41, 0.97)	.037
Meeting MET min/wk guidelines (n = 1296)		0.53 (0.34, 0.80)	.003		0.65 (0.42, 1.01)	.06
Meeting pedometer steps/d guidelines (n = 1081)		1.02 (0.59, 1.77)	.94		1.27 (0.70, 2.29)	.43

Notes. BMI = body mass index (weight in kilograms divided by the square of height in meters); CI = confidence interval; MET = metabolic equivalent of a task; MVPA = moderate to vigorous physical activity; OR = odds ratio. Physical activity was the main outcome, and health literacy (reference: adequate) was the main explanatory variable.

^aAdjusted for age, gender, education, income, employment, ethnicity, smoking status, BMI, and number of chronic conditions.

In the adjusted analysis, we did not observe any significant associations between HL and self-reported physical activity for both MVPA (b = −57.8; 95% CI = –132.1, 16.6; P = .128) and MET minutes (b = −218.3; 95% CI = –502.4, 65.9; P = .132; Table 3). However, HL was significantly associated with achieving guidelines for self-reported physical activity, whereby those with inadequate HL had a lower likelihood for achieving guidelines for self-reported MVPA (OR = 0.63; 95% CI = 0.41, 0.97) and MET minutes (OR = 0.65; 95% CI = 0.42, 1.01). Although the association with MET minutes was marginally significant, the point estimate was suggestive of a lower likelihood for achieving guidelines. Similar to unadjusted results, there were no statistically significant associations between HL and pedometer-measured walking (b = 39.0; 95% CI = -667.7, 745.8; P = .914) or achieving guidelines for this measurement based on a suggested threshold of 8000 or more steps per day (OR = 1.27; 95% CI = 0.70, 2.29). Results from linear regression were also consistent with those from γ regression analysis (MVPA minutes per week: exponentiated coefficient = 0.73; 95% CI = 0.53, 0.99; P = .05; MET minutes per week: exponentiated coefficient = 0.74; 95% CI = 0.54, 1.01; P = .06; pedometer steps: exponentiated coefficient = 1.02; 95% CI = 0.88, 1.18; P = .84).

DISCUSSION

In this population-based sample, we found that approximately 1 in 10 older adults have inadequate HL, a prevalence much lower than what has been observed in the general population in Alberta (F. Al Sayah, written communication, January 2015), or in other populations.¹³ We also found that a small proportion of older adults achieved guidelines for MVPA physical activity, with significant variations between those with adequate and inadequate HL. Based on self-reported physical activity, older adults with inadequate HL were less likely to achieve guidelines based on recommended thresholds. Nonetheless, such differences were not observed for pedometer-measured walking, and the results of this study should be interpreted in light of the findings.

Our results add to the current literature examining the relationship between HL and self-reported health behaviors. For example, other reports suggest that older adults with inadequate HL are more likely to have a sedentary lifestyle than those with adequate HL,¹⁶ and that inadequate HL is associated with poor compliance with guidelines for physical activity.¹⁹ However, our self-report findings and those of others suggesting that older adults with inadequate HL are less likely to achieve physical activity guidelines are contradictory to one study that found nonsignificant associations between physical activity and HL in an adult population.²¹ In addition, a more recent longitudinal study reported a significant association between inadequate HL and poor and declining physical functioning among older adults.³⁵ The discrepancies among these studies might be explained by differences in study populations and methods for assessing the exposure and outcome variables, but a common characteristic among them was the use of different self-reported measures of physical activity, which may have yielded unreliable measurements.³⁶ Despite these discrepancies, our results should be interpreted in light of our secondary finding.

The secondary aim of our study was to investigate the association between HL and objectively measured physical activity. Perhaps the most interesting finding in this study is the difference in the results obtained from self-reported physical activity compared with those from pedometer-measured walking. Previous studies suggest that the measurement method might have a significant impact on the observed levels of physical activity in all population groups including older adults.^36,37 Self-report measures of physical activity were reported to be both higher and lower than objective estimates of physical activity, which poses a problem for both reliance on self-report measures and for attempts to correct for differences between self-report and objective measurements.³⁷ Furthermore, the correlation between subjective and objective measures of physical activity was reported to be moderate (r = 0.38) in a systematic review that included 36 studies on physical activity measurement in older adults.³⁶

Our study provides evidence that this discrepancy between subjective and objective measurements of physical activity leads to inconsistent research outcomes, and limits the interpretability and implications of findings. Furthermore, as there were no differences in pedometer-measured walking between the HL groups, these results suggest the possibility of an effect of HL on the reporting of physical activity measured by self-report questionnaires. Future research in this area should consider the use of more comprehensive objective measures of physical activity such as accelerometers.

Aside from the inherent limitations of self-report measures, the physiological and cognitive changes that occur with aging make accurate measurement of physical activity and other outcome measures more challenging in this population. Older adults learn new information at a slower rate than younger adults because of a decline in fluid intelligence, defined as the reasoning and processing components of learning, which makes mental multitasking and comprehending the abstract more challenging.³⁸ Previous research suggests that a declining cognitive function in older adults is associated with a decline in HL abilities,³⁸ and subsequently with poor performance of health care tasks.³⁹ This might explain the results observed in this study, whereby participants with inadequate HL might have inaccurately underreported their actual physical activity compared with those with adequate HL skills. This is likely a result of the evident mismatch between individuals’ HL skills and the health management demands placed upon them by health systems, particularly for those with chronic health needs,^40,41 resulting in literacy-based barriers to good health. The improvement of individuals’ HL skills and reaching a balance between the level of these skills and the health care system literacy demands is a public health imperative, and should be a major goal of health organizations in Canada in light of the adverse effects of inadequate HL, not only on health behaviors, but also on morbidity and mortality.^14,17,18

Limitations

Although this study involved a large sample of older adults and used subjective and objective measures of physical activity, the results should be interpreted in light of a few limitations. First, although pedometers are considered an objective measure of walking and superior to self-report, it is important to note that they provide limited information about the type of physical activity, and do not capture certain activities such as swimming, resistance exercise, upper-body movements, cycling, or complex movements. However, because physical activity in this age group is generally preferred of a moderate level such as walking,⁴² it is prudent to assume that the pedometer captured the majority of daily physical activity among the participants.

Second, some of the measures used in this analysis (e.g., BMI, number of chronic conditions) were based on self-report, which inherently have some limitations and might lead to bias. In this study, an underreporting of weight, height, and number of chronic conditions would likely be the case, and would have led to an underestimation of the impact of these factors.

Third, although we adjusted for several sociodemographic and health-related characteristics that are known to be associated with physical activity, other factors such as mobility restrictions (e.g., hip surgery), walkability of neighborhoods where participants reside, and availability of walking spaces during the winter were not accounted for in this analysis.

Finally, despite the study sampling approach that aimed to recruit a widely representative sample, the majority of participants in this study were White older adults with high education and low levels of inadequate HL, and thus the results of this study may not be generalizable to other groups of older adults, particularly those from minority populations in which the prevalence of inadequate HL is much higher.

Conclusions

In this study, we found that inadequate HL was associated with less likelihood of meeting guidelines for self-reported physical activity; however, these associations were nonexistent when we used an objective measure of physical activity. These results affirm that inconsistencies in physical activity research outcomes may be attributable to the method for data collection (i.e., self-report vs objective, and measurement tools). The results also raise the question of whether HL affects the accuracy of respondents’ reporting on self-report questionnaires. In light of the difficulty of accurate measurement of physical activity in older adults because of the physiological and cognitive changes that occur with aging, we suggest that the studies examining physical activity in this population should consider the use of both objective and self-report measures.