The Inventory of Physical Activity Barriers for Adults 50 Years and Older: Refinement and Validation

Abstract

Background and Objectives

Due to health consequences associated with insufficient physical activity (PA), particularly among aging adults, health care providers should assess and address lack of PA participation. Addressing lack of PA means developing individualized PA prescriptions that incorporate solutions to PA participation barriers. Assessing PA participation barriers can be done through the Social Ecological Model-based Inventory of Physical Activity Barriers Scale (IPAB). This study aimed to refine the initial 40-item IPAB and determine its reliability and validity.

Research Design and Methods

Five hundred and three community-dwelling adults 50 years and older completed a demographic and health questionnaire, the Physical Activity Vital Sign, the IPAB, and a feedback questionnaire. For scale refinement, half of the data were used for exploratory factor analysis and the other half for confirmatory factor analysis. The refined scale underwent reliability and validity assessment, including internal consistency, test–retest reliability, and construct validity.

Results

The refined scale contains 27 items consisting of 7 factors and 1 stand-alone item: (a) environmental, (b) physical health, (c) PA-related motivation, (d) emotional health, (e) time, (f) skills, (g) social, and (h) energy (a stand-alone item). The 27-item IPAB has good internal consistency (alpha = 0.91) and high test–retest reliability (intraclass correlation coefficient = 0.99). The IPAB’s mean scores were statistically different between those who met the recommended levels of PA and those who did not (p < .001).

Discussion and Implications

The information gathered through the IPAB can guide discussions related to PA participation barriers and develop individualized PA prescriptions that incorporate solutions to the identified barriers.

Background and Objectives

Approximately 30% of adults older than the age of 50 and 35% of adults older than the age of 75 report no leisure-time physical activity (PA), leading to significant health-related consequences such as increased risk of injury, frailty, falls, disability, morbidity, and mortality (Green & Fielding, 2011; Watson et al., 2016). These consequences can lead to decreased quality of life and an increased need for caregivers, hospitalizations, and institutionalizations, resulting in significant caregiver and national health care burden (Green & Fielding, 2011). The long-term repercussions associated with insufficient PA levels highlight the need for health care providers to address the current lack of PA in aging populations (Lobelo et al., 2014; Lobelo & de Quevedo, 2016).

One way to address insufficient levels of PA is by assessing and addressing patients’ PA participation barriers (Hillsdon et al., 2005). We developed a PA participation barrier scale called the Inventory of Physical Activity Barriers (IPAB) to assist health care providers in identifying barriers that can be used to formulate solutions that address PA participation barriers. Due to the complexity and multifactorial nature of insufficient PA, we developed the IPAB based on the Social Ecological Model (Wingood et al., 2021). The Social Ecological Model helps to elucidate the complex interplay between individual, social, institutional, community, and public policy factors that affect behavior (McLeroy et al., 1988). Individual factors are any characteristics of the individual and include a person’s knowledge, attitudes, behavior, self-concept, health, and skills (McLeroy et al., 1988). Social factors are related to formal or informal social networks or support systems, such as family, work, workout partners, and friends (McLeroy et al., 1988). Institutional factors are related to social institutions with organizational characteristics and formal or informal rules and regulations for operations (McLeroy et al., 1988). Examples of institutions and organizations include community organizations, neighborhoods, and independent or assisted living facilities. Community factors include relationships among organizations and informal networks within defined boundaries (McLeroy et al., 1988). Examples of community factors include cultural values, norms, and build environment. And public policy factors are related to local, state, and federal public policies and laws that affect the decision to perform a behavior (McLeroy et al., 1988).

The purpose of the IPAB is to give health care providers a tool that helps organize their PA participation barrier assessment and guide treatment. Health care providers can use the IPAB’s results to initiate conversations with patients about identified PA participation barriers. Furthermore, this information can be used to develop individualized PA prescriptions that incorporate solutions to PA participation barriers, an evidence-based approach of increasing PA levels among aging adults (Thornton et al., 2016). Health care providers can also administer the IPAB pre- and postintervention to assess the change in amount or magnitude of PA participation barriers. The purpose of this study was to refine the Social Ecological Model-based IPAB scale and determine the refined scale’s reliability and validity.

Research Design and Methods

Study Design

We used survey data collected between June and December 2020 to refine the IPAB and explore the scale’s psychometric properties, including internal consistency, contrasting group validity, and dimensionality. Most participants completed a one-time survey. A randomly selected subsample from the survey study completed the IPAB twice, approximately 2 weeks apart, to assess test–retest reliability. Our University’s Internal Review Board (IRB) approved the study. All participants completed the IRB-approved consenting process before initiating data collection. The Checklist for Reporting Results of Internet E-Surveys was used to guide both phases of this study (Eysenbach, 2004).

Participants

Participants were community-dwelling adults who reported being at least 50 years old and able to read, comprehend, and write in English. We excluded individuals who needed assistance from another person to leave their house or lived at an assisted living or long-term care facility. Power calculations revealed that factor analysis required the largest number of participants to reach optimal power, resulting in a sample size of ≥500 participants, with half of the data needed for exploratory factor analysis (EFA) and the other half for confirmatory factor analysis (CFA; Fabrigar et al., 1999). For test–retest reliability, a total of 50 participants were needed to complete the IPAB twice (Park et al., 2018). Due to potential attrition rates between the first and second survey, 80 participants were recruited to complete the IPAB twice for the test–retest reliability analyses (Park et al., 2018). Participation was voluntary and all participants were entered into a raffle drawing to win $25.00.

Recruitment

Recruitment strategies included targeted advertisements, social media messages, listserv postings, flyers, and e-mails to health care providers and agencies that serve older adults (e.g., physical therapists, physicians, senior centers, assisted living facilities, and Area Agencies on Aging). Additional recruitment occurred using the Claude D. Pepper Older Americans Independence Center registry (IRB0503150) at the University of Pittsburgh. All potential participants received an e-mail with the study link to REDCap (Vanderbilt University, Nashville, TN), a secure data collection software.

Data Sources

Brief questionnaire

Participants completed an online questionnaire asking them about the following: (a) age, (b) gender, (c) race and ethnicity, (d) education, (e) income, (f) marital status, (g) rurality, (h) ability to walk or wheel a half a mile, (i) 12-month retrospective history of falls, (j) height and weight, (k) if they like PA, (l) if they would like to be more physically active, (m) if they think that participating in higher amounts of PA would be beneficial, (n) if they think PA is important, (o) if a health care provider has told them to be more physically active, and (p) if conditions or diagnosis related to their muscles or bone, heart or lungs, sensory system, mental or cognition, and degenerative disorders or diagnosis are barriers to their PA. The questions were based on several previously published questionnaires, including Census (Census, 2010), National Health and Aging Trends Study (NHATS, 2019), and National Health and Nutrition Examination Survey (Centers for Disease Control and Prevention, 2020). For additional details, see Supplementary Material for a copy of the questionnaire.

Physical activity vital sign

The Physical Activity Vital Sign (PAVS) is a reliable and valid two-question scale that asks individuals about their weekly moderate-to-vigorous level of PA (Greenwood et al., 2010). Compared to accelerometer-measured PA, the PAVS correctly identifies individuals who meet the weekly recommended 150 min of moderate-to-vigorous intensity PA (κ = 0.46, p < .05; Ball et al., 2015).

The IPAB

The IPAB is a 40-item scale that assesses PA barriers among individuals aged 50 years and older (Wingood et al., 2021). Scale respondents are informed that the purpose of the scale is to “learn about your barriers to doing PA at a moderate-to-vigorous intensity.” The IPAB then provides respondents with the following definition of moderate-to-vigorous PA: “Physical activity is any movement that is produced by your muscles and requires energy. It is classified as a moderate intensity when you can talk, but not sing during the activity and a vigorous activity is when you cannot complete full sentences during the activity. Examples include gardening, active house chores, brisk walking, climbing stairs, biking, or dancing” (Wingood et al., 2021). Respondents then complete the 40 items, which all have the same stem: “My physical activity is limited, because …” Response options are on a 5-point Likert scale (“Never” [1], “Rarely” [2], “Sometimes” [3], “Often” [4], and “Always” [5]; Wingood et al., 2021). Items are averaged to create a mean score with averages closer to five indicating more barriers (Wingood et al., 2021). According to pilot data, the 40-item IPAB can differentiate between individuals who did and did not meet 150 min/week of moderate-to-vigorous PA (p = .01) and had an internal consistency of α = 0.97 (Wingood et al., 2021).

Feedback questionnaire

After completing the IPAB, all participants received a feedback questionnaire. The questionnaire asked about ease of completion, length of time, clarity of the questions, and their perception of the IPAB’s acceptability and appropriateness. The data were collected using a 5-point Likert scale (“Completely Disagree” [1], “Disagree” [2], “Neither Disagree/Agree” [3], “Agree” [4], and “Completely Agree” [5]) asking participants to rate the level of agreement with each item. Participants were asked to provide a reason for all statements that they disagreed with.

Procedures

All participants received a link to REDCap where they completed the brief questionnaire, PAVS, IPAB, and a feedback questionnaire. Each questionnaire was provided on a separate page. For test–retest reliability, a random subgroup of 80 participants were asked to repeat the IPAB about 10–14 days after the initial data collection. Only completed surveys were used for data analysis.

Analysis

We computed descriptive statistics and tests of normality (Kolmogorov–Smirnov) on all the variables. Two items (“it’s hard to see the benefit of PA” and “others have told me to avoid PA”) were identified as being skewed and therefore they were eliminated from the scale (Clark & Watson, 1995). All other items were identified to be normally distributed. Scale refinement, exploration of IPAB’s dimensionality, and examination of the scale’s validity were assessed by running an EFA on a randomly selected half of the sample. To ensure sampling adequacy, we completed the Bartlett test and Kaiser–Meyer–Olkin test of sampling adequacy (Yong & Pearce, 2013). Because we eliminated the two skewed items and all other data were normally distributed, we performed our EFAs using a maximum likelihood estimation (Fabrigar et al., 1999). Secondary to previous research and our preliminary analysis identifying that factors will correlate with each other, we used an oblique rotation (Promax; Briggs & Cheek, 1986). The number of factors was determined using eigenvalues >1.0 (Yong & Pearce, 2013). Items within each factor were eliminated if they did not have a factor loading of at least 0.30 (Yong & Pearce, 2013). We then identified items as cross-loading if they rounded up to 0.30 on additional factors and therefore were also eliminated (Costello & Osborne, 2005; Matsunaga, 2010; Tabachnick et al., 2007).

The EFA results were confirmed by running CFA on the other half of the sample (Cole, 1987). For the CFA, the estimation methodology was maximum likelihood. We explored our CFA’s goodness of fit by examining the chi-square test of perfect fit, root-mean-square error of approximation, comparative fit index, Tucker–Lewis index, and standardized root-mean-square residual (Cole, 1987).

The refined 27-item IPAB’s internal consistency was determined using Cronbach alpha (Yurdugul, 2008). Intraclass correlation coefficient (ICC 2,k) was used to assess test–retest reliability (Koo & Li, 2016; Park et al., 2018). As there is no gold standard for assessing PA participation barriers, we did not look at criterion-related validity. Instead, because PA levels are inversely associated with the amount of PA participation (Jones & Nies, 1996; Reichert et al., 2007), we examined the IPAB’s construct validity by using a t-test and Cohen’s d to compare the IPAB’s mean scores of participants who met the recommended 150 min of weekly moderate-to-vigorous PA and those who did not.

Results

Participant Characteristics

Out of the 508 individuals who completed the consent form, 503 completed the entire study, resulting in a 99% completion rate. Our 503 participants had a mean age of 70.1 (SD = 8.5) years and were primarily women (69.9%), non-Hispanic White (95.4%), had at least a college degree (81.3%), had an annual income greater than $45,000 (77.8%), were married (65.6%), with more than half meeting the recommended levels of PA (59.3%). For additional details, see Table 1. The participants had an average IPAB score of 1.8 of 4.0 (SD = 0.5) with a range of 1.0–3.2. See Supplementary Table 1 for descriptive statistics related to individual items on the IPAB.

Table 1.

Participants’ Demographics and Sample Characteristics

Variable	Mean (SD)or n (%)
Age, mean, years (SD)	70.0 (8.5)
Women, n (%)	351 (69.8)
Race or ethnicity, n (%)
Non-Hispanic White	478 (95.4)
Black or African American	11 (2.2)
Asian	4 (0.8)
Middle Eastern or North African	1 (0.2)
Latino/a	4 (0.8)
Some other race/ethnicity	4 (0.8)
Education level, n (%)
College degree or higher	409 (81.3)
Annual household income, n (%)
<$45,000	112 (22.3)
$45,000–100,000	202 (40.2)
>$100,000	189 (37.9)
Marital status, n (%)  Married	331 (65.7)
Rurality, n (%)
Rural/country	79 (15.7)
Suburban	267 (53.1)
Urban/city	157 (31.2)
Ability to walk or wheel half a mile, n (%)	472 (93.6)
One-year retrospective history of fall, n (%)	116 (23.1)
Body mass index, mean (SD)	27.5 (9.3)
Complete ≥150 min of moderate–vigorous weekly PA, n (%)	298 (59.3)
Like PA, n (%)	430 (85.5)
Reports wanting to be more physically active, n (%)	432 (85.9)
Feels that more physical activity would be helpful, n (%)	447 (88.9)
Perceive physical activity as an important topic, n (%)	445 (96.7)
Health care provider has told them to be physically active, n (%)	258 (51.3)
Reports the following health conditions as barriers to PA, n (%)
Muscles- or bones-related conditions	214 (42.6)
Heart- or lungs-related conditions	58 (11.5)
Sensory systems-related conditions	17 (3.4)
Mental or cognition-related conditions	17 (3.4)
Degenerative disorders or diagnosis	37 (7.6)
Enjoy using a computer, n (%)	385 (83.7)

Note: PA = physical activity.

Dimensionality and Item Reduction

After eliminating the two items (“it’s hard to see the benefit of PA” and “others have told me to avoid PA”) with a skewness or kurtosis greater than ±2.5, we randomized our sample into two groups: the EFA (n = 251) was conducted using data from Group 1, and the CFA (n = 252) was conducted using data from Group 2. The two subsamples did not have any significant differences (p > .05) in demographic or potential confounders listed in Table 1. Based on the Kaiser–Meyer–Olkin test (0.89) and Bartlett’s test of sphericity (chi-squared = 4,333; df = 703; p < .001), we had sampling adequacy and were able to perform our EFA and CFA.

After several iterations of the EFA, we eliminated 11 items with low factor loading (<0.3) or cross-loading (factor loading of ≥0.3 on two or more factors). During the EFA analysis, we identified that the last factor had two potential items. However, one of these items “Of the fatigue I feel from my daily demands” was identified to cross-load and therefore was deleted, resulting in a one-factor solution. Because “I don’t have enough energy” had a factor loading of 1.1, it was decided to keep the item as a “stand-alone” item. Our decision to keep it as a stand-alone item was further supported by the item’s mean, which was higher than any other item (2.7/5.0; SD = 0.8) and corrected item-total score correlation of 0.6. The Bartlett’s test (p < .001) and Kaiser–Meyer–Olkin (0.9) were re-run with the remaining 26 items and indicated sampling adequacy. The results identified the following additional seven factors with an eigenvalue ≥1.0: (a) environmental (six items), (b) physical health (five items), (c) PA-related motivation (four items), (d) emotional health (four items), (e) external factors (three items), (f) skills (two items), and (g) social (two items). The two-item factors were kept due to support provided by the Social Ecological Model and previous research (Ayotte et al., 2010; Bethancourt et al., 2014). See Supplementary Table 1 for a detailed view of the final eight factors, items within each factor, items removed from the scale, and the items’ factor loading. Our seven factors had a common variance of 64.2.

To validate the EFA model, we conducted a CFA on the second half of the sample (n = 252). See Figure 1 for a visual representation of the CFA. Our chi-square was high (406.2) and significant (p < .001). Given that chi-square has multiple limitations, including the large impact of sample size, a high chi-square was expected (Cole, 1987; Hu & Bentler, 1998). The comparative fit index and Tucker–Lewis index were above the recommended ≥0.90 (comparative fit index = 0.94, Tucker–Lewis index = 0.93; Schumacker & Lomax, 2004). The standardized root-mean-square residual measure was 0.05, which is below the recommended fit of 0.08, indicating a good fit (Hu & Bentler, 1999). The good fit was further confirmed with the root-mean-square error of approximation (0.05) being below 0.08 and having a p value of .53 (MacCallum et al., 1996).

Figure 1.

Confirmatory factor analysis. PA = physical activity.

Reliability

Cronbach’s alpha for the final scale was 0.91, indicating the scale is internally consistent but not overly redundant (Tavakol & Dennick, 2011). When looking at the reliability of each factor, Cronbach’s alpha was (a) 0.82 for environmental, (b) 0.81 for physical health, (c) 0.85 for PA-related motivation, (d) 0.76 for emotional health, (e) 0.67 for external factors, (f) 0.69 for skills, and (g) 0.68 for social. Test–retest reliability, measured by ICC (2,k), was 0.99 for the total score, indicating that the IPAB has excellent test–retest reliability (Koo & Li, 2016).

Construct Validity

Participants who completed less than 150 min of moderate-to-vigorous PA had an average IPAB score of 2.1 (SD = 0.4) and a range of 1.0–3.2, whereas participants who met PA recommendation levels had an average score of 1.6 (SD = 0.4) and a range of 1.0–2.9. The statistical difference between these means (p < .001) indicates that the IPAB was able to differentiate between individuals who meet the recommended levels of moderate-to-vigorous PA and those who did not, illustrating construct validity. The results are further supported by the effect size (Cohen’s d = 0.38; Gignac & Szodorai, 2016).

Feedback

Out of the 503 participants, 460 participants provided feedback. The initial 40-item IPAB took participants an average of 9.0 (SD = 4.6) minutes to complete. The majority of the participants agreed that the IPAB meets their approval (78.9%) and stated that they would do it if their health care provider gave it to them (95.0%). The participants stated that it was easy to complete (93.4%), directions were clear (92.4%), it was a good length (90.7%), it was beneficial (63.3%), it was an important topic (95.7%), and identified the scale as respectful (92.6%). Participants’ feedback is presented in further detail in Table 2. Participants who did not find the scale easy to use, an appropriate length, or important provided the following feedback: (a) “if given by health care provider it would be nice to have a link to complete at home”; (b) “some questions are hard to answer, wish there was a not applicable”; (c) “need space to explain problems”; and (d) “lack of clarity about what a ‘scale’ is.”

Table 2.

Survey Feedback

Variable	Mean (SD)or n (%)
Time, average minute (SD)	9.0 (4.6)
Meets my approval, n (%)
Agree–completely agree	365 (79.4)
Neither agree/disagree	9 (2.0)
Disagree–completely disagree	86 (18.7)
It is easy to use, n (%)
Agree–completely agree	434 (94.4)
Neither agree/disagree	15 (3.3)
Disagree–completely disagree	12 (2.6)
The directions are clear, n (%)
Agree–completely agree	426 (92.6)
Neither agree/disagree	19 (4.1)
Disagree–completely disagree	13 (2.8)
It is a good length, n (%)
Agree–completely agree	418 (91.0)
Neither agree/disagree	38 (8.3)
Disagree–completely disagree	4 (0.9)
It is beneficial, n (%)
Agree–completely agree	298 (64.8)
Neither agree/disagree	158 (34.4)
Disagree–completely disagree	4 (0.9)
It is respectful, n (%)
Agree–completely agree	416 (90.4)
Neither agree/disagree	22 (4.8)
Disagree–completely disagree	4 (0.9)
If my health care provider gave me this, I would complete it, n (%)
Agree–completely agree	441 (95.9)
Neither agree/disagree	13 (2.8)
Disagree–completely disagree	6 (1.1)

Discussion

Using EFA and CFA, the IPAB was revised to a 27-item scale with the following seven factors and one stand-alone item: (a) environmental, (b) physical health, (c) PA-related motivation, (d) emotional health, (e) time, (f) skills, (g) social, and (h) energy (stand-alone item). The psychometric evaluation of the 27-item IPAB identified a high internal consistency, test–retest reliability, and construct validity.

Prior to developing and validating the IPAB, options to assess factors affecting PA participation included incomplete scales (Brown, 2005), scales validated on specific populations such as individuals with mobility impairments (Becker et al., 1991; Rimmer et al., 2004; Vasudevan et al., 2015), or a compilation of multiple unifactorial scales. Using multiple unifactorial scales means using separate scales to comprehend the personal, social, and community factors that affect PA participation. Examples of unidimensional scales include personal scales (e.g., the Physical Activity Enjoyment Scale [Mullen et al., 2011], Outcome Expectation Scale [Resnick, 2005], Index of Self-Regulation [Yeom et al., 2011], and Multidimensional Self-Efficacy for Exercise Scale [Rodgers et al., 2008]), social support scales (e.g., the Physical Activity Social Support Assessment Scale [Reis et al., 2011]), and environmental factor scales (e.g., the Neighborhood Environment Walkability Scale [Cerin et al., 2006]). Collating results from multiple scales is time-consuming, not clinically feasible, and can result in health care providers feeling overwhelmed, particularly when considering the other responsibilities health care providers have (Yamada et al., 2015). Additionally, the majority of scales have been designed to assess factors affecting PA participation within a single level of the Social Ecological Model rather than across individual, social, environmental, and policy levels. An alternative method of assessing factors affecting PA participation is asking an open-ended question about the factors keeping adults 50 years and older from being physically active. Using an open-ended question also has multiple limitations, including lack of completeness, a greater amount of time needed to obtain information about PA participation factors (Glaab et al., 2010), the potential of biasing the patient toward a more favorable response, and the inability to quantify PA participation factors or assess the impact of an intervention addressing PA participation factors (Jette et al., 2009).

Ecological Momentary Assessment (EMA) is an alternative, dynamic method of assessing factors affecting PA participation, including PA barriers and expectations in a day-to-day or moment-to-moment scenario (Cain et al., 2009). The EMA method has been identified as useful for identifying decreased self-efficacy for PA during increased episodes of pain (Sperber et al., 2014) and factors related to intentions to participate in PA (Conroy et al., 2011). The greatest benefits of the EMA are the ability to gather information about these dynamic PA participation factors and the reduction of recall bias (Cain et al., 2009). However, EMA is not always feasible or acceptable (Trull & Ebner-Priemer, 2009), and assessment questionnaires like the IPAB could be useful when there is insufficient time to complete an EMA. Because none of the methods discussed are ideal within the health care setting, we developed the IPAB to give health care providers a needed, feasible, self-administered, multifactorial assessment tool of PA participation barriers.

Using the self-administered IPAB takes less than 10 min to complete and provides health care providers with an in-depth understanding of their patients’ PA participation barriers. Health care providers can use the IPAB’s findings to guide their conversations to address identified barriers and develop an individualized PA prescription. For example, if the PA participation barriers are the fear of being injured or it is hard to find places to be physically active, health care providers can educate patients about a safe PA that does not require a specific location. By incorporating this information, health care providers have the potential of empowering their patients and increasing the patients’ PA level (Hillsdon et al., 2005). According to the Health Belief Model, the impact of addressing PA participation barriers is even greater among individuals who do not have sufficient self-efficacy to engage in PA or do not perceive that the PA benefits outweigh the barrier(s) (Champion & Skinner, 2008). Therefore, if a barrier affects an individual’s self-efficacy, the value of addressing the PA participation barrier is even greater (Jancey et al., 2009).

A major strength of our study is that the IPAB is the first multifactorial assessment tool that has been validated for PA participation barriers among community-dwelling adults aged 50 years and older. The other major strength is that the eight factors identified by the EFA and confirmed by the CFA align with our foundational theoretical basis, the social ecological model. More specifically, the individual factors of the socioecological model are assessed by the physical health, emotional health, PA-related motivation, time, skills, and fatigue factors of the IPAB. The social factors of the socioecological model that affect PA participation are assessed by the social factor of the IPAB. And the community- and policy-level factors of the socioecological model are assessed by the environmental factor of the IPAB.

Prior to implementing the IPAB, several study limitations must be considered. Our primary limitation is related to the lack of diversity in the participant sample with respect to race/ethnicity, education, and PA participation levels. The lack of diversity among our participants may result in challenges with the generalizability of the current study findings. For example, according to Mathews et al. (2010), American Indians are more likely to identify build environment and lack of knowledge about PA as barriers to participating in regular PA (Mathews et al., 2010). Therefore, if we had more American Indians participate in our study, the mean score of PA participation barriers related to environment and knowledge may have had a higher average score. Another example is related to education. Having a higher level of education may have affected the results because older adults who have a higher level of education are more likely to meet the recommended amount of PA (Ashe et al., 2009), and meeting the recommended levels of PA affects the amount of PA participation barriers a participant may have (Reichert et al., 2007). Thus, it is essential to note that without additional research with a more diverse sample, we are unable to determine if the factor structure may be invariant. A minor limitation that may explain the above-average levels of PA is the use of a subjective PA assessment scale (PAVS) as epidemiological studies have identified that PA assessments that are based on self-report questionnaires can result in overreporting (Matthews et al., 2012). Additionally, our participants had an above-average enjoyment in using computers, which may have skewed the feedback regarding ease of survey completion (Levine et al., 2016). It is also important to note that participant feedback was based on the initial 40-item scale.

Conclusions

We have refined and validated the previously published IPAB, a questionnaire that health care providers can use to assess PA participation barriers and develop individualized PA interventions to increased PA participation among patients 50 years and older. The 27-item refined scale is internally consistent, has high test–retest reliability, and can differentiate between individuals who met the recommended levels of PA and those who did not. The IPAB can be used to guide health care providers through a conversation about PA participation barriers, develop individualized PA prescriptions that incorporate solutions to the identified barriers, and examine the effectiveness of implemented solutions.

Funding

Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health (P20GM135007 to Vermont Center for Cardiovascular and Brain Health [D. M. Peters]). This work was also supported in part by the National Institute on Aging of the National Institutes of Health (K24 AG057728 and P30 AG024827 to J. S. Brach).

Conflict of Interest

None declared.