| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Thomas R. Prohaska, School of Public Health, University of Illinois at Chicago, 1603 W. Taylor St., Chicago, IL 60612. E-mail: prohaska{at}uic.edu
 |
Abstract |
|---|
 TOP Abstract BackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
Key Words: Exercise • Retention • African American • Chronic illness • Functional status
There is an increasing emphasis on expanding outreach of health research and health promotion programs to traditionally underrepresented older populations including disadvantaged populations and minority and nonminority older adults with chronic illnesses and disability. This is particularly true for community-based exercise health promotion programs with older adults. Findings from community-based exercise interventions have been reported with older African Americans (Prohaska, Peters, & Warren, 2000; Yanek, Becker, Moy, Gittelsohn, & Koffman, 2001) and older populations with chronic illnesses and disability (Jette et al., 1999; Schmidt, Gruman, King, & Wolfson, 2000). Despite the growing attention to targeting underrepresented older groups for exercise health promotion programs, there is a paucity of information on the effectiveness of various recruitment strategies and program intervention approaches on rates of recruitment and exercise maintenance once recruited to exercise programs. This study evaluates how recruitment strategies and program characteristics in an exercise intervention study influence enrollment and retention of African American and White older adults with multiple chronic illnesses.
Prohaska, Walcott-McQuigg, Peters, and Li (2000) noted that a lack of participation in exercise programs by older adults can be examined at two broad levels: (a) attrition during recruitment and (b) program attrition or relapse among those who were successfully recruited into the exercise program. King and colleagues (1992) organized the factors affecting recruitment attrition and exercise program attrition into three broad categories: person-based factors, environmental factors, and program intervention factors. Person-based factors include demographic characteristics of older persons and their perceptions and beliefs about exercise. Environmental influences include the program setting as well as interpersonal resources influencing participation (e.g., social support, physician recommendations). Intervention factors include study design characteristics such as recruitment strategies, inclusion/exclusion criteria, and definitions of compliance and participation.
Much of the research on the factors associated with exercise participation in older adults has focused on the determinants of retention rather than recruitment into the exercise programs. Similarly, examination of the factors associated with recruitment and retention of older adults in exercise programs has focused primarily on person-based characteristics of the older population and environmental influences. With some noted exceptions, relatively less attention has been directed toward understanding how intervention factors influence recruitment and maintenance of exercise among older adults (Mills, Stewart, King, Roitz, & Sepsis, 1996; Mills et al., 2001). Characteristics of the recruitment strategy and the exercise intervention may be the most influential factors affecting participation in exercise among traditionally underrepresented older populations (Wagner, Grothaus, Hecht, & LaCroix, 1991). This study pays particular attention to the programmatic decisions that influence who is recruited into, and who is retained in, an exercise program.
|
Background |
|---|
|
TOPAbstractBackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
Perceived benefits and barriers are also well-documented person-based and environmental factors known to influence recruitment and maintenance in exercise programs among older adults. Perceptions of self-efficacy in performing the exercise activities are associated with a greater likelihood of adoption and maintenance of regular exercise (McAuley, Lox, & Duncan, 1993; Prohaska et al., 2000; Sharpe & Connell, 1992). Barriers such as a lack of time, employment, and care-giving commitments have also been associated with nonparticipation in regular exercise among older adults (Lee, 1993; Marcus, Rakowski, & Rossi, 1992; Prohaska et al., 2000).
Program Intervention Factors
The influence of program intervention factors on recruitment of older adults into exercise programs has been reported in King, Harris, and Haskell (1994) and McNeely and Clements (1994). These studies have supported the general conclusion that a broad recruitment strategy (e.g., media, announcements, letters to older community residents) without prescreening will result in a large number of potential participants being exposed to the recruitment message. However, such recruitment procedures also result in a small proportion of participants being eligible to be enrolled (McNeilly et al., 2000). Halbert, Silagy, Finucane, Withers, and Hamdorf (1999) noted that out of 2,878 letters sent to potential participants, only 299 older adults met the inclusion criteria and agreed to participate in an exercise advice program. Alternatively, targeted recruitment procedures—such as physician referral; presentations at senior housing sites, nutrition sites, or retirement homes; or mailed letters to Medicare-eligible seniors—result in greater initial effort for a small pool of participants but a higher probability of identifying eligible participants.
However, the effect of such recruitment strategies may not be identical across diverse older populations. Issues such as exposure to the recruitment message and ability to read the recruitment materials may create a selection bias against those with low education or who have literacy difficulties. Similarly, some screening tools for cognitive impairments, such as the Mini-Mental State Examination, are affected by low literacy and education levels (Albert & Teresi, 1999; Ostrosky-Solis, Lopez-Arango, & Ardila, 2000). This can result in excluding older adults with low education as being cognitively impaired (Mayeaux et al., 1995) and may inappropriately exclude potential participants.
The decision to use physician referral versus self-referral is another potentially important program factor that may influence participation among underrepresented populations. Typically, physician approval is required for most exercise health-promotion research interventions with older populations. Partnering with community physicians for referral of older participants can be a valuable and cost-effective recruitment tool (Hirsch et al., 1992). However, it assumes that all older adults have a regular physician and that the physician is motivated to make referrals to the exercise program. It is possible that these assumptions are questionable in underrepresented older populations.
In addition, the more invasive the screening procedure in terms of medical tests, the less likely it is that certain underrepresented groups will even have access or agree to participate in the screening process. In particular, many African Americans have an extended history of distrust of medical and public health research (Gamble, 1997). Research looking at the effect of how the screening protocol and specific inclusion and exclusion criteria influence the enrollment of participants has not been examined.
Another factor is the health status of the older participant. Most exercise programs for older adults exclude persons with health problems (Buchner et al., 1993; King, Haskell, Taylor, Kraemer, & DeBusk, 1991). As noted above, older adults with poor health are more likely to self-select out of participation in exercise programs. It is not known how inclusion criteria requiring participants to have multiple chronic illnesses influence decisions to participate in an exercise program.
Finally, the definition of who constitutes an older adult is often based on chronological age, which may bias self-selection into the program. Many adults may not perceive themselves as being older and may thus not respond to recruitment messages directed to an older population. Yet we know there can be significant differences mentally, physically, and emotionally between the young-old and the oldest old (Marín et al., 1995).
With respect to retention, the most crucial programmatic factor is the definition of study compliance and which participants are classified as dropouts. This is a critical decision for longitudinal studies with multiple data collection points. However, even short-term studies can suffer from high attrition depending on the criteria established. The determination of what data collection method to invoke can also influence compliance. For example, face-to-face interviews in a central location may have reduced results due to transportation problems, but mailed, written survey instruments may have reduced response in a low-literacy population.
In addition, factors related to the exercise program conduct could be crucial to an individual's continued participation; the frequency, intensity, and duration of the exercise program must be appropriate to participants' health and fitness ability. The friendliness of the instructor, consistency of instruction, and sometimes the race or ethnicity of the instructor, the research staff, and the other participants may be an issue.
Environmental Factors
Neighborhood safety has been positively associated with exercise participation (Centers for Disease Control and Prevention, 1996). Participants desiring to exercise on their own may be impeded from walking because their neighborhood is unsafe on account of gang activity, broken sidewalks, no sidewalks, or unshovelled sidewalks (Clark, 1999). Convenience in the form of logistics may seem minor to researchers but be of great importance to participants. The nearness of the site to transportation, the hours of availability, and the cost can all be environmental factors that may form a significant psychological barrier to the exerciser. "It's just too far, I can't get on the machines I want, the adult lap swim isn't at a good time for me, there are no lockers where I can store things between visits."
To summarize, this study examines how recruitment and retention of underrepresented populations, specifically minority and nonminority older adults, is associated with study intervention factors, including physician screening and referral, the use of media and other recruitment strategies, the inclusion criteria of multiple chronic illnesses, age criteria, and exercise location. The data analyzed here represent 273 responses and referrals to a variety of recruitment mechanisms for an exercise intervention for older adults with multiple chronic illnesses. Subsequently, 103 participants were enrolled (completed a baseline interview and fitness performance tests). This analysis examines how decisions about recruitment strategy and program content and logistics influenced the eligibility of potential participants and ultimately the retention of our sample.
|
Design and Methods |
|---|
|
TOPAbstractBackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
The first intervention site was a new neighborhood child and family center located within two blocks of the community clinic. The facility is unused during the day when children are in school and contains the necessary resources for an exercise program for older adults: large gymnasium, lockers, bleachers, water fountains, and a parking lot and is located just one block from a major bus line. The facility also contains a satellite clinic run by Dr. David Freedman that is used for assessments, blood pressure checks, and measuring height and weight. Environment and participant safety were key concerns as participants were chosen because of their illness status and because the intervention was being conducted in a community facility instead of a closely monitored hospital or research setting. These factors also influenced our choice of instructor. The instructor, a White man, is certified in cardiopulmonary resuscitation and has experience working with older adults in cardiac rehab programs.
All written materials (recruitment flyers, instrument instructions, informed consent documents) were composed at an 8th-grade reading level. Baseline and follow-up assessments were conducted face to face at the intervention site by trained interviewers (75% African American) in order to accommodate low reading or literacy levels in the sample. The informed consent process was approved by the university Office of Protection of Research Subjects.
Interested participants were contacted by phone, at which time additional screening was conducted. Participants who passed the initial screening and were interested in participating in the study were randomized to either a home-based exercise control group or a group-based exercise and education intervention. All participants had a baseline assessment conducted after signing an informed consent document. Each assessment lasted from 1 to 2 hr. Participants in the group-based program started exercising at the next available session following their baseline interview.
Participants in the group-based exercise intervention meet 2 days per week for 45 min of exercise and 15 min of health education about how exercise can be used to help manage chronic disease. The exercise intervention meets continuously for the duration of the study with breaks for holidays. Participants are encouraged to attend for the entirety of their enrollment (2.5 years). The exercise session consists of gentle warmup exercises done in chairs, gradual progression of aerobic walking, strengthening exercises using therabands, and then T'ai Chi Chuan to cool down. Exercises are performed to big band and gospel music. The education component focuses on the physiology of the four chronic diseases, their symptoms and treatments, and how exercise could positively affect their chronic illnesses and symptom management. The education information is delivered in a group discussion format and participants are encouraged to bring up other topics that they are interested in learning about. Participants' blood pressure is taken before and after each class. There are two class sessions at each site with approximately 5–15 participants attending at any given session.
Each participant randomized to the home-based group received an exercise manual and theraband equipment with personalized instruction from the exercise leader after completing the baseline evaluation. They were instructed to exercise on their own at least two times per week at a comfortable intensity level. They received a phone call from the exercise leader about 3 weeks after receiving their materials to assess how they were doing, ask if they were using the materials, and to modify the program if there were problems.
Follow-up assessments are conducted for all participants at 3 months, 6 months, and every 6 months thereafter for 2.5 years. Each participant is given $5 for every completed interview (seven interviews total). Participants are sent birthday cards, get well cards, seasonal holiday cards, and letters with updates on study findings. At significant milestone assessments, participants are rewarded with program incentives such as T-shirts, water bottles, socks, and exercise videos to use at home. The study also paid for participants to attend annual fitness events sponsored by the city Department of Aging.
Revised Design
The initial recruiting strategy using physician referrals resulted in only 29 possible participants (24 of whom enrolled). Interviews with clinic physicians revealed that they perceived that their low-income patients would spend their limited resources on transportation to the exercise program instead of going to doctor's appointments; thus only three physicians participated and sent referrals. Based on this, we decided to continue the physician recruitment process but expand the recruitment strategy to include targeted presentations at senior housing sites, city aging activities, and local churches. We also broadcast public service announcements on radio and cable access television and placed advertisements in local newspapers.
This change to the recruitment protocol required us to change our screening process and determination of eligibility. With direct recruiting, we now depended on participants' self-reports of the presence or absence of their chronic illnesses. An additional step was added to obtain their physician's permission for them to participate in the exercise program as well as to confirm their chronic illnesses and determine any exclusion criteria. In addition to widening our recruitment net, we also lowered our age requirement to 50 years. On the basis of initial screenings, there were many African Americans who met the chronic illness criteria but were not yet 60 years old.
In Year 2 we added a second site to increase our overall rate of recruitment. This second site is in the same city in a location that contains a large Eastern European and Greek population. Participants at this church-based site were recruited through local churches, social workers in senior housing buildings, and a partnership with a community aging foundation. The same exercise instructor conducted the intervention at both sites.
Some participants refused to enroll in the program because of how they were randomized. Some preferred to exercise at home because if they had to get up and physically go to another location, they knew that they "just wouldn't go." Others felt strongly that they needed to exercise with a group for motivation and they wanted the discipline of a set schedule. Rather than lose these participants during recruitment, our protocol was to randomize people and encourage them to participate in their assigned groups but allow them to participate in their preferred setting.
|
Measures |
|---|
|
TOPAbstractBackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
Independent variables related to person-based factors are from participants' phone screen and baseline interviews. The baseline instrument assessed psychosocial measures of cognitive impairment using the Mini-Mental State Examination1 and health distress, psychological well-being and energy/fatigue were measured using the Medical Outcomes Study approach (Lorig et al., 1996). The last three measures had Cronbach's alphas of.82,.84, and.88, respectively. Overall self-ratings of health were assessed from the SF-36 Health Survey (McDowell & Newell, 1996), using the question "In general, would you say your health is..." with responses ranging from 1 = excellent to 5 = poor. Participants completed the short form of the Geriatric Depression Scale (McDowell & Newell, 1996) and a measure of self-efficacy to exercise regularly (Lorig et al., 1996), with a scale ranging from 1= not confident to 10 = totally confident.
Physical functioning (across multiple domains) was assessed using a modified version of the Stanford Health Assessment Questionnaire Disability Scale (Lorig et al., 1996; McDowell & Newell, 1996) and was reliable for our sample (
=.93). Fitness performance measures (6-min walk [Rikli & Jones, 1999a, 1999b] and modified step test) were conducted to assess endurance and aerobic ability.
Risk factors were assessed through a checklist of chronic diseases and a checklist of 21 possible symptoms associated with the four chronic diseases. Participants who stated that they experienced a particular disease or symptom were then asked to assess how much difficulty the disease or symptom caused them (on a scale ranging from 0 = no difficulty to 3 = unable to do). A maximum illness difficulty score and a maximum symptom difficulty score were derived along with a simple count of the number of illnesses and symptoms. Body mass index was calculated from participants' weight and height.
Programmatic variables are primarily related to the inclusion and exclusion criteria: age, physician confirmation of chronic illness, and decision to allow participants to refuse randomization. Attendance at the class is taken regularly. As most participants were sedentary at baseline and the overall fitness level of the sample was low, we did not specify any requirements for intensity. The perceived frailty of the cohort indicated that using traditional measures of intensity would not be meaningful for this sample.
|
Results |
|---|
|
TOPAbstractBackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
Person-Based Factors Related to Recruitment
Basic demographic characteristics for all self-reported eligibles were collected during the phone screen. A comparison of enrollees and nonenrollees can be seen in Table 1. Our referrals were predominantly female and not currently married and over half are African American. Of the enrollees, about 20% did not graduate from high school and 30% earn less than $10,000 per year. Individuals who did not enroll were younger, t (189) = -1.979, p =.049; more likely to be under 60, 
2 (1, N = 191) = 3.87, p =.036; and more likely to self-report having diabetes, 2 (1, N = 190) = 3.93, p =.033. Among enrollees, African Americans were less likely than Whites to be male, 2 (1, N = 103) = 6.55, p =.01; to have a high school education, 2 (1, N = 102) = 5.74, p =.015; or to earn $10,000 or more, 2 (1, N = 95) = 5.03, p =.021.
|
|
In terms of other risk factors, we required participants to have two of the four chronic conditions to be eligible; almost all participants (83%) reported additional chronic health problems. Other health issues reported included depression (43%); cancer (5%); lung diseases such as emphysema, tuberculosis, or asthma (16%); ulcers (16%); stroke or neurological problems such as Parkinson's disease (11%); and circulatory problems (45%). Participants reported experiencing an average of almost six symptoms during the past week. Excess weight affected a majority of our sample. Overall, 22% of the sample is overweight (body mass index > 25 and < 30), and 68% are obese (body mass index > 30). Although only 8% currently smoke, half of the remainder are former smokers with an average smoking duration of 20 years (data not shown).
Program Intervention Factors Related to Recruitment
Our decision to lower the age requirement from 60 to 50 increased our number of referrals by 26% (70). Those younger than age 60 composed 25% of self-reported eligibles and made up 19% of enrollees. The assumption that this would significantly increase the number of African Americans who would be referred and enrolled was not proven. Both African Americans and Whites under 60 who were eligible enrolled in equal numbers.
Owing to the change in the recruitment protocol to do direct recruiting, we had the opportunity to compare participants' self-reports of chronic illness with their physician's reports of chronic illnesses, as shown in Table 3. For all chronic illnesses, there were significant discrepancies between the percentage of people who believed they had a chronic illness and the percentage of people whom the physicians reported had that chronic illness. Half of the enrolled participants had at least one discrepancy with their physician regarding the presence or absence of a chronic illness. In particular, participants overestimated the presence of arthritis, with 23% believing that they had this condition. Although having any disparity across the four illnesses was unrelated to race, gender, or self-ratings of overall health, race was related to the disparity with hypertension. Whites were more likely to underestimate the presence of hypertension, 2 (1, N = 95) = 4.21, p =.04. In general, discrepancies were due to participants' overestimating the number of chronic illnesses they had.
|
2 (1, N = 103) = 10.63, p =.002, and to be currently married, 2 (1, N = 103) = 4.8, p =.028. The assumption that these people would not have enrolled if they had not been permitted to choose their groups cannot be tested.
Environmental Factors Related to Recruitment
The decision to add a second site had major consequences for the overall sample. We were unable to find a setting in a similar demographic community that had equivalent facilities, availability, safety, and accessibility to public transportation. Our decision to be consistent in environmental factors caused a significant shift in our sample in terms of both demographics and health measures. The sample from the first site was more likely to be African American, 2 (1, N = 103) = 35.96, p <.000; to not have completed high school, 2 (1, N = 102) = 7.58, p =.004; and to have an income less than $10,000, 2 (1, N = 95) = 4.98, p =.002. Participants from the first site reported having less energy, t (98) = -2.46, p =.016; had more difficulty with eating, t (99) = 2.60, p =.011; more difficulty with reaching, t (99) = 2.03, p =.045; less difficulty with hygiene, t (99) = -2.11, p =.040; and experienced more limits on their social activities because of their health, t (99) = 2.79, p =.006.
Factors Related to Retention
Attrition and retention factors are presented in Table 4. Overall attrition was highest in the first 3 months of the intervention program with 21% (22) of the sample dropping out between baseline and the 3-month assessment. Dropout status is not officially confirmed until 1 year after the last contact (either class participation or data collection point). Another 7% of the sample dropped between 3 months and 6 months and an additional 2% dropped between 6 months and 12 months. Thus, at 1 year, study retention was 70% with 31 official dropouts.
|
The physical health extremes of our participants played a role in retention (see Table 4). Participants who rated their health as poor or fair at baseline were significantly more likely to drop out in the first 3 months of the program, 2 (1, N = 103) = 5.51, p =.018. Although those in poor or fair health were significantly more likely to have worse psychosocial scores at baseline, psychosocial scores were not related to dropout at either the 3-month assessment or at 1 year. Self-efficacy to exercise regularly was significant at 1 year but still very high, demonstrating the ceiling effect for this measure. Functional abilities, as demonstrated by 6-min walk results and activities of daily living (ADL) subscale measures, and difficulties experienced from chronic illness and chronic illness symptoms differentiated dropouts from retainees at 1 year.
The maximum ADL difficulty score demonstrates the broad spectrum of functional abilities in our sample. A full 18% of participants reported no difficulty performing any of 22 ADL or Instrumental ADL items. At the opposite end, 29% reported a maximum ADL difficulty score of 3 ("unable to do") for at least one item. Participants who were the least functionally capable were significantly more likely to drop out before 3 months, 2 (1, N = 101) = 3.85, p =.048, and at 1 year, 2 (1, N = 101) = 14.92, p <.000, than participants who scored in the no difficulty, mild, or moderate categories on maximum ADL difficulty.
Although no systematic interviews with dropouts were conducted, the instructor recorded observations and updates on participants in field notes. Of the total dropouts who were assigned to the group exercise program (n = 22), 10 never began the exercise program—effectively dropping out just after the baseline interview. Of the 12 who started attending the exercise class, 58% dropped out because of poor health.
Program Intervention Factors Related to Retention
Of the participants who started the group-based program, participation in the form of attendance was quite high. In the first 3 months, 89% of the treatment group attended 75% or more class sessions. Between 3 and 6 months, 85% of participants had similar attendance and another 8% attended fewer than 50% of the sessions. Between Months 6 and 12, 69% attended three fourths of the sessions. Those who came to the exercise class initially generally had good attendance. Thus, attrition does not appear to be related to program content or the exercise instructor. Those participants who chose their exercise setting were no more likely to be retained than those who were randomized (based on chi-square tests). There were no significant differences between the treatment group and the home-based group.
|
Discussion |
|---|
|
TOPAbstractBackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
Our specific attempts to programmatically reduce attrition of African Americans during recruitment and during the intervention were generally successful. When group-specific efforts are used to recruit and retain African Americans, they are ultimately no more likely to drop out than Whites. Using African American research staff, recruiting in African American settings, providing facilities in the local community, and tailoring the program content were effective (Marín et al., 1995). However, we were less successful at retaining participants disadvantaged by lack of education. Having a high school degree was significant for retention at both 3 months and 1 year. We should note that African Americans composed 75% of those who did not graduate from high school; thus, they were at greater risk of dropping out. Lack of education could be related to retention in several ways: Participants with less education may have felt overwhelmed by the length of the interview, the detailed informed consent process, or the complexities of the standardized scales on the baseline instrument. Although we tried to reduce the dependency on reading ability by having face-to-face interviews, we were not entirely successful in eliminating educational issues. Persons under 60 were also less likely to enroll. As this group is not officially retired, they may have had other work opportunities (Prohaska et al., 2000).
Clinical research studies depend on confirmed biomarkers of disease. In retrospect, this study could have recruited participants on the basis of their perceptions of having chronic illnesses. Illness perceptions are potentially more relevant to the adoption of a particular self-care behavior such as exercise. Had we relied solely on self-reports of chronic illness, our sample could have increased by 83% if we had enrolled all those persons who believed themselves eligible (n = 191). Previous research evaluating self-reports of chronic conditions has indicated that older adults typically underreport their ailments (Porell & Miltiades, 2002). This finding was not supported in our study, although it should be noted that it benefited individuals to overstate their health conditions. If underreporting is the norm, it is possible that had we used physician verification for all referrals, a large number of referrals who self-reported being ineligible because they did not have two chronic conditions might in fact have been eligible. Thus, the determination of health or illness status should be considered carefully, depending on the goals of the study.
The extremes of health that composed our sample complicated the exercise program's conduct and content. Although our intention was to target individuals with clinical disease who could use exercise as a way to manage their illness, many participants had so many limitations that it was logistically difficult for them to attend. Thus, functional limitations and interference from illness symptoms was significantly related to both early and later dropout. It is challenging to design a program intense enough for the healthy participants that also accommodates those participants with functional limitations and assistive devices.
Although the physicians were originally concerned about the environmental factor of transportation costs, cost did not appear to be an issue. Only five self-reported eligibles indicated that they did not have transportation to the exercise site. Transportation was problematic, but transportation cost was not cited as a reason for attrition. Participants who reported problems with transportation were those persons who used the public transportation special cars and vans provided for disabled persons.
No systematic data on environmental factors are available. Some individuals indicated that they would not participate because of the neighborhood for the first site. One participant did not like that the facility was located near two schools. Another did not like that the site was on a cul-de-sac that she termed a dead-end street. This site did experience significantly more dropouts at 1 year. These anecdotal comments highlight the importance of perceived neighborhood safety to older adults.
With a continual exercise program of this duration (2.5 years), we could define dropout very broadly. This allowed us to retain many participants who would not ordinarily be considered as participating. Thus, vacationers were welcomed back, as were participants who returned after surgery, caregiving responsibilities, death of a family member, or other health or family concerns. Alternatively, the length of the study may have influenced some potential participants to not enroll.
Just over 20% of our sample dropped in the first 3 months during the adoption phase (Schmidt et al., 2000). This is considerably less than other studies looking at attrition in frail older adults, where 36% dropped out in the first 3 months (Schmidt et al., 2000). Participants who drop out during this time period may be less exercise nonadopters as they are program misfits. The temptation is to conclude that these early dropouts are individuals who are not ready to adopt the behavior. As this analysis indicates, numerous program decisions can affect recruitment and retention. Early dropouts may try the program but find it inconvenient, have had different expectations about what the exercise program would be like, or not like the instructor or other participants. Attrition in this time period may be less about the actual behavior and more about program details and the fit with the participant's capabilities. Randomization may also influence early dropout. There may have been participants who agreed to randomization even though they had a preference, and this may have contributed to their attrition. Randomization has been shown to be a barrier to recruitment in exercise studies (Yanek et al., 2001), but little is known about its effect on study attrition.
|
Implications |
|---|
|
TOPAbstractBackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
This study is limited as it did not include a complete analysis of environmental factors because the two sites had significantly different locations, neighborhoods, and populations. Although site was a significant factor in dropout at 1 year, it is unclear whether it was due to environment or a significant difference in person-based factors between sites.
|
Footnotes |
|---|
1 School of Public Health, University of Illinois at Chicago, Chicago, IL.
2 Mile Square Health Center, University of Illinois at Chicago, Chicago, IL.
A cutpoint of 18 was used to denote cognitive impairment. Participants who met this criteria participate in the program but have limited data items.
Received for publication July 11, 2002. Accepted for publication September 16, 2002.
|
References |
|---|
|
TOPAbstractBackgroundDesign and MethodsMeasuresResultsDiscussionImplicationsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  D. Liu and N. G. Castle Health Information Technology in Nursing Homes Journal of Applied Gerontology, February 1, 2009; 28(1): 38 - 58. [Abstract] [PDF]
|
|
|
|
 |
|
 D. Burns, A. C. M. Soward, A. H. Skelly, J. Leeman, and J. Carlson Effective Recruitment and Retention Strategies for Older Members of Rural Minorities The Diabetes Educator, November 1, 2008; 34(6): 1045 - 1052. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Resnik, D. Liu, D. L Hart, and V. Mor Benchmarking Physical Therapy Clinic Performance: Statistical Methods to Enhance Internal Validity When Using Observational Data Physical Therapy, September 1, 2008; 88(9): 1078 - 1087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. G. Castle and J. Engberg Further Examination of the Influence of Caregiver Staffing Levels on Nursing Home Quality Gerontologist, August 1, 2008; 48(4): 464 - 476. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. M. Bellows and H. A. Halpin MDS-Based State Medicaid Reimbursement and the ADL-Decline Quality Indicator Gerontologist, June 1, 2008; 48(3): 324 - 329. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Jancey, A. Clarke, P. A. Howat, A. H. Lee, T. Shilton, and J. Fisher A Physical Activity Program to Mobilize Older People: A Practical and Sustainable Approach Gerontologist, April 1, 2008; 48(2): 251 - 257. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. G. Castle, J. Engberg, and Aiju Men Nurse Aide Agency Staffing and Quality of Care in Nursing Homes Med Care Res Rev, April 1, 2008; 65(2): 232 - 252. [Abstract] [PDF]
|
|
|
|
 |
|
 E. A. Miller and V. Mor Balancing Regulatory Controls and Incentives: Toward Smarter and More Transparent Oversight in Long-Term Care Journal of Health Politics Policy and Law, January 1, 2008; 33(2): 249 - 279. [Abstract] [PDF]
|
|
|
|
|
|
G. Arling, R. L. Kane, C. Mueller, J. Bershadsky, and H. B. Degenholtz Nursing Effort and Quality of Care for Nursing Home Residents Gerontologist, October 1, 2007; 47(5): 672 - 682. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. D. Phillips, R. Shen, M. Chen, and M. Sherman Evaluating Nursing Home Performance Indicators: An Illustration Exploring the Impact of Facilities on ADL Change Gerontologist, October 1, 2007; 47(5): 683 - 689. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. G Castle, J. Engberg, and D. Liu Have Nursing Home Compare quality measure scores changed over time in response to competition? Qual. Saf. Health Care, June 1, 2007; 16(3): 185 - 191. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Kane, G. Arling, C. Mueller, R. Held, and V. Cooke A Quality-Based Payment Strategy for Nursing Home Care in Minnesota Gerontologist, February 1, 2007; 47(1): 108 - 115. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. Knox, E. C. Blankmeyer, and J. R. Stutzman Comparative Performance and Quality Among Nonprofit Nursing Facilities in Texas Nonprofit and Voluntary Sector Quarterly, December 1, 2006; 35(4): 631 - 667. [Abstract] [PDF]
|
|
|
|
|
|
D. G. Stevenson Is a public reporting approach appropriate for nursing home care? Journal of Health Politics Policy and Law, August 1, 2006; 31(4): 773 - 810. [Abstract] [PDF]
|
|
|
|
|
|
C. Harrington, J. O'Meara, M. Kitchener, L. P. Simon, and J. F. Schnelle Designing a Report Card for Nursing Facilities: What Information Is Needed and Why Gerontologist, April 1, 2003; 43(90002): 47 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Czaja and J. Sharit Practically Relevant Research: Capturing Real World Tasks, Environments, and Outcomes Gerontologist, March 1, 2003; 43(90001): 9 - 18. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
