PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of gerontLink to Publisher's site
 
Gerontologist. 2011 December; 51(6): 867–875.
Published online 2011 June 30. doi:  10.1093/geront/gnr066
PMCID: PMC3220665

Resident-Directed Long-Term Care: Staff Provision of Choice During Morning Care

Sandra F. Simmons, PhD,1,2,* Annie Rahman, MSW,3 Linda Beuscher, GNP, PhD,1,4 Victoria Jani, DC,1 Daniel W. Durkin, PhD,1 and John F. Schnelle, PhD1,2

Abstract

Purpose: To develop an observational protocol to assess the quality of staff–resident communication relevant to choice and describe staff–resident interactions as preliminary evidence of the usefulness of the tool to assess current nursing home practices related to offering choice during morning care provision. Design and Methods: This study included 73 long-stay residents in 2 facilities. Research staff conducted observations for 4 consecutive morning hours during targeted care activities (transfer out of bed, incontinence, dressing, and dining location). Observations were conducted weekly for 12 consecutive weeks. Staff–resident interactions were measured related to staff offers of choice and residents’ responses. Results: Interrater agreement was achieved for measures of staff offers of choice (kappa = .83, p < .001), type of choice provided (kappa = .75, p < .001), and resident requests related to choice (kappa = .72, p < .001). Observations over 2,766 care episodes during 4 aspects of morning care showed that staff offered residents choice during 18% of the episodes. Most observations (70%) were coded as staff offering “no choice.” Implications: Nursing home staff can use a simplified version of this standardized observational tool to reliably measure staff–resident interactions related to choice during morning care provision as a first step toward improving resident-directed care practice.

Keywords: Long-term care, Quality of care, Measurement, Quality of life, Assessment of conditions/people, Consumer-directed care

In a no-nonsense manual on implementing culture change in nursing homes, Bowers, Nolet, Roberts, and Esmond (2007) asked providers, “Is your facility providing person-centered care? Are you unsure if you are really providing it? What does person-centered care actually mean?” These are timely questions, for there is growing support for resident-directed care in nursing homes, with an emphasis on offering residents choices.

The rationale for this focus is rooted in Ryan and Deci’s (2000) self-determination theory, for resident choice, viewed as an expression of one’s autonomy, is a central concept within this life-span theory. Specifically, this theory posits that three universal psychological needs—for competence, relatedness, and autonomy—must be met for individuals to enjoy an optimal sense of personal well-being. The theory also holds that a person’s social environment can foster attainment of these needs, such that the person thrives, or the environment can thwart these needs, to the person’s detriment. Thus, for example, nursing homes that elicit and honor resident choices help to foster residents’ sense of autonomy and, with it, their sense of well-being. Failure to attend to choice may undermine residents’ autonomy, leading to such feelings as passivity and hopelessness.

In recent years, new nursing home policies and practices have emerged in line with the tenets of self-determination theory. Federal regulatory guidelines, for example, now identify choice over daily schedules as a resident right (Centers for Medicare and Medicaid Services [CMS], 2009). These guidelines instruct surveyors to ask residents if they are offered choices about daily life activities, such as bedtimes and dining, and observe whether those choices are honored by staff. Federal and state surveyors responsible for ensuring nursing home care quality can issue care deficiencies, which require a formal plan of correction from the nursing home, as part of their survey process if they determine that choice is not routinely offered to residents. Similarly, popular culture change models (e.g., Eden Alternative, Green House, Wellspring) emphasize the importance of personal choice and freedom (Bowers et al., 2007), and nursing home staff have self-reported they offer residents daily care choices such as when to get up, when to bathe, and what to eat (Doty, Koren, & Sturla, 2008). Few studies, however, have collected independent data to confirm these reports or published assessment tools that could help providers answer Bowers and colleagues’ (2007) pointed question: “Is your facility providing person-centered care?” The lack of objective data and assessment tools also has limited efforts to evaluate the impact of resident-directed care on clinical and quality-of-life outcomes. Thus, the purpose of this study was twofold: (a) to develop an observational protocol to assess the quality of staff–resident communication relevant to choice and (b) to describe staff–resident interactions in a small sample as preliminary evidence of the usefulness of the tool to assess current nursing home practices related to offering choice during morning care provision.

Conceptual Approach

Researchers and experts in a wide range of fields, including health care, emphasize the value of using reliable accurate data to drive improvement efforts for two reasons. First, such data facilitates the replication of evidence-based practices (e.g., Deming, 1986; Pronovost & Vohr, 2010). This purpose may be especially important in promoting resident-directed care in nursing homes. Presently, as Bowers and colleagues (2007) have observed, “there is very little useful documentation available about how culture-change initiatives have actually been implemented and achieved. Vague suggestions … are not useful to organizations wanting to replicate the changes. What does one actually do [to achieve the desired outcomes]?” (p. 9). Data-driven assessments can help answer this question.

Second, a large body of research has demonstrated the value of using reliable accurate data to guide improvement initiatives, evaluate and monitor program results, and inspire workers to sustain efforts that show positive outcomes (Deming, 1986; Neily et al., 2010; Pronovost et al., 2006). Nursing home culture change advocates have echoed these calls for accurate data collection to drive quality improvement efforts (Bowers et al., 2007; Quality Partners of Rhode Island, 2011). Similarly, they have cautioned nursing home staff against accepting “impressions as evidence of … achievement of outcomes” (Bowers et al., 2007, p. 105), noting that “evidence collected to confirm these impressions can surprise everyone” (Bowers et al., 2007, p. 105). Such may be the case with resident choice. In a recent study, researchers conducted observations of daily care interactions between staff and residents in 20 facilities across five states (Schnelle et al., 2009a, 2009b). In all 20 facilities, staff did not offer residents choices in at least one of the observed daily care areas (Schnelle et al., 2009b).

Accurate measurement, however, requires appropriate tools and, at present, the data collection tools available to assess resident choice in nursing homes are inadequate to the task. In concept, such a tool should directly address the central question posed by Bowers and colleagues (2007): “In a facility where they have fully integrated person-centered care, no decision is made without asking: ‘Has the resident participated in this decision?’” (Bowers et al., 2007, p. 12). Although the newly revised Minimum Data Set (MDS) assessment requires staff to ask residents about the importance of making choices related to daily schedules, including bedtimes and what to wear, these assessments are conducted only at admission and annually thereafter (CMS, 2010). Many resident choices, however, must be made daily or on a per-care-episode basis. For example, what time a resident prefers to get up, what he/she wants to wear, or where he/she prefers to dine may change from day to day. In short, it is insufficient to assess residents’ preferences at one point in time for many aspects of daily care. Instead, staff should communicate with residents about their preferences as part of their daily interactions around care provision to truly provide resident-directed care.

The primary purpose of this study was to develop and test a standardized tool that researchers, surveyors, and nursing home staff could use to reliably assess staff–resident communications relevant to choice—a first step toward improving resident-directed care practice. The study focuses on four specific aspects of morning care: when to get out of bed, incontinence care, what to wear, and where to have breakfast. These four activities were targeted because they typically occur together within a predictable timeframe, so there are multiple opportunities to offer residents choices on a daily basis. However, the measurement principles used in this study apply generally to the provision of choice in other care areas. As a secondary goal, the study describes the use of the tool to assess staff–resident interactions for a sample of residents as preliminary evidence of current nursing home care practices during morning care provision.

Methods

Participants and Setting

Participants were recruited from two facilities housing a total of 300 residents. Nurse aide-to-resident ratios, as reported by administrative staff, were similar between the two sites and averaged 8.3 residents to one aide on the day shift, 11.5 residents to one aide on the evening shift, and 16.2 residents to one aide on the night shift. Nurse aide staff was not assigned to the same residents each day in either of the two sites. Upper-level staff at both facilities self-reported that they had completed previous staff training about the importance of offering residents choices, in keeping with recent culture change initiatives. Neither facility had survey citations in this area (one facility had a “5-star” CMS rating).

A total of 175 residents met study inclusion criteria, which required residents to be long stay (non-Medicare) and able to respond to simple yes/no questions during a screening interview. Written consent was obtained from the resident or designated proxy for 79 (45%) of the 175 eligible residents. The university-affiliated institutional review board approved the consent procedures. A total of 73 residents completed the observational data collection during a 3-month period under usual care conditions.

Measures

Demographic information was retrieved from each participant’s medical record in addition to his/her most recent MDS 2.0 assessment. An MDS-derived measure of physical functioning was calculated based on seven items (Morris, Fries & Morris, 1991), yielding scores ranging from 0 (independent in all areas) to 28 (completely dependent in all areas). The most recent MDS and care plan were reviewed to assess staff documentation of residents’ daily care preferences related to morning care activities. Cognitive status was assessed with the Mini-Mental State Examination (MMSE), with a score range from 0 (severely cognitively impaired) to 30 (cognitively intact; Molloy, Alemayehu, & Roberts, 1991).

Observations of Morning Care

Research staff observations targeted four morning care activities: transfer out of bed, incontinence care (changing and/or toileting), dressing (what to wear), and breakfast dining location. Trained researchers conducted continuous observations for an average of 3.5 hr (up to four continuous hours) per resident during weekdays (Monday to Friday) in each facility. The goal was to observe each participant at least once per week (minimum of 4 hr on 1 week day) throughout the 12 study weeks, with the weekday of observation varying each week for the participant. The observation period was adjusted at each site (6–10 a.m. or 7–11 a.m.) based on the morning care routine.

Staff Communication Relevant to Choice

Standardized observations during daily care provision in our previous studies led to the reliable coding of three mutually exclusive types of staff prompts that reflected different levels of encouragement for residents to make a choice (Schnelle et al., 2009a, 2009b). These staff prompts were active choice, passive choice, and no choice. Active choice prompts encouraged the resident to make a specific decision. For example, “Do you want to get up now or after breakfast?” Passive choice prompts required the resident to only assent to care, but in practice the care routine was often initiated before the resident made a response. For example, “It’s time to get up now, okay?”

There were three subcategories within the “no choice” category: (a) staff did not provide care or speak to the resident, (b) staff provided care without conversation, and (c) staff provided care with conversation. In the first subcategory, staff did not enter the resident’s room at any point during the continuous 4-hr observation period (6–10 a.m. or 7–11 a.m.); thus, the resident was not given an opportunity to either receive care or make a choice about care. In the second, “no conversation” subcategory, the staff member said nothing at all to the resident during care provision. In the third subcategory (care with conversation), conversation was defined as the staff member informing the resident of the care being provided but without any option for choice or assent, such as “It’s time to get up.” This category also included staff conversation unrelated to the care routine (e.g., “Good morning. How are you today?”). Conversation that occurred between staff and residents without care provision or in the context of care that was not the focus of this study (e.g., medication pass) was not counted in the observational data.

Resident Responses to Care Delivery

If care was provided without choice (no-choice category), research staff documented whether the resident (a) assented to the care (e.g., replied “Okay”), (b) requested something different (e.g., “Get me up later”), or (c) complied with the care activity without verbally providing assent or expressing an alternative preference (e.g., followed staff instructions without comment). Research staff recorded all verbal responses to prompts and care verbatim.

Residents’ Spontaneous Requests for Care

On some occasions, residents were observed to spontaneously request care or otherwise make their preference clear before staff prompted them or provided care. These occasions were coded as “spontaneous requests” for care and may also have included a staff response (i.e., active, passive, or no choice) such that coding for spontaneous requests were not mutually exclusive from coding for staff prompts. For example, if a resident pressed their call light and told staff upon entering the room and without any prompting from staff, “I want to get up now,” then a “spontaneous request” would be counted for the “transfer out of bed” care activity. Both a spontaneous request and an active staff prompt would be coded if a staff member asked the resident what he/she wanted in response to a call light (e.g., “What do you need?”) before the resident made his/her request.

Staff Responses to Resident Requests

When residents expressed a choice about care, either spontaneously or in response to a staff prompt, research staff documented whether the resident’s stated choice was honored by staff within 5 min of the request. A 5-min criterion was set to ensure timeliness of staff response to resident requests, with the exception of dining location, which was dependent upon time of breakfast meal service.

Reliability of Coding

Despite the extensive preliminary work in the initial development of the observational protocol, it required 2 weeks at the beginning of this study to refine coding definitions due to unanticipated scenarios. Once the coding definitions were further refined, however, approximately 3 hr of training was sufficient to achieve agreement. Interrater agreement was determined based on two research staff observing the same resident and care episode while independently coding the observation. Research staff were considered trained and their data reliable only after significant kappa agreement was achieved. These kappa values ranged from .78 to −.84 prior to the start of data collection for this study. The project coordinator continued to conduct interrater reliability checks twice per month with each observer to prevent observer drift during the 3 months of data collection.

The average kappa value across the four care activities for the presence or absence of any type of choice was .83 (n = 142 observations, p < .001). For type of choice offered by staff (active vs. passive), the average kappa value was .75 (n = 158 observations, p < .001). The kappa value for whether residents made a spontaneous request related to choice averaged .72 (n = 130 observations, p < .001).

Results

Participants and Setting

Participants were predominately women and White with an average age of 83.7 years and an average length of residency of 3.5 years (Table 1). They were moderately cognitively impaired as indicated by an average MMSE total score of 14.5. Participants were rated by facility staff on their most recent MDS 2.0 assessment as moderately physically dependent with an average MDS–activities of daily living total score of 17.5. There was no chart documentation in either the MDS 2.0 or the care plan of participants’ daily care preferences related to the four morning care activities that were the focus of this study. Most notably, there was no documentation that any of the participants were bed bound or otherwise preferred to remain in bed until late in the morning.

Table 1.
Characteristics of Participants (N = 73)

Staff Communication Relevant to Choice

Table 2 shows the observational data for the 73 participants during the four targeted morning care activities across all 12 study weeks. There were a total of 2,766 observations across all four care activities (see Table 2, last column). On average, each participant was observed 1.17 (±0.16) days per study week, or 3.5 total hours each observation day. Each 3.5-hr observation period yielded a total of four data points per person per week (one data point for each of the four care activities), regardless of whether the care activities occurred together during the same observation period. When care was provided, staff provided more than one care activity 66% of the time within the same observation period (e.g., resident was assisted out of bed, to the toilet and dressed). However, the occurrence or nonoccurrence of each care activity was counted separately in the data because a resident could receive some aspects of care and not others and/or be offered choice in some care domains but not others.

Table 2.
Observations of Staff Offers of Choice During Morning Care

Table 2 shows the proportion of observations that each type of communication occurred within each care area and the total across all four care areas. The number of observations differed by care area due to variability in staff care provision. The first row of Table 2 shows the proportion of observations during which research staff directly observed the participant to perform the task independently (without staff assistance). This proportion ranged from 4% to 9% across all four care areas. Participants who completed the care activity independently were not rated on staff provision of choice (i.e., these cases were excluded from subsequent rows in Table 2). The second row of Table 2 shows the proportion of observations during which care was provided but the communication between staff and residents was not directly observed, most often due to the care being provided prior to the beginning of the observation period (e.g., resident was already out of bed and dressed prior to 6 a.m.). This proportion ranged from 4% to 22% across the four care areas, and these cases also were not rated on staff provision of choice.

For the remaining observations during which care was provided and observed in each of the four care areas, staff offered choice during 18% of the observations overall (Table 2, “Total” column). Active choice was offered during 14% of all observations, with active prompts most commonly offered for getting out of bed (20%) and dressing (18%). The remaining 4% of cases overall were instances wherein staff offered passive choice (i.e., resident assent).

The majority of observations (70%) were coded as staff offering “no choice” (Table 2, “Total no choice offered”). Within the three subcategories of “no choice offered,” the most common occurrences were “provided care with conversation” (50% of the “no choice offered” category) and “no care or conversation observed” (43% of the “no choice offered” category). As previously described (see Methods), “provided care with conversation” was defined as the staff member informing the resident of the care being provided but without any option for choice or assent. This was most common in the “dining location” area (Table 2, 83% of the “no choice offered”) with comments such as “It’s time for breakfast.” The other category, “no care or conversation observed,” meant that research staff did not observe the care activity to occur or any interaction between the staff and the resident at any point during the continuous 4-hr morning observation period. This was least likely to occur for “dining location” (4%) because most residents were served breakfast, even if they remained in bed, relative to the other three care areas (range 56%–61%). To ensure the accuracy of these data, the observation time period was adjusted (from 6–10 a.m. to 7–11 a.m.) and repeated observations were conducted. These adjustments did not result in fewer observations of “no care or conversation observed.”

Participants were observed to make spontaneous requests for assistance during 6% of all observations (Table 2, “Total” column, last row). Most spontaneous requests were for toileting assistance and incontinence care (11%). Usually, spontaneous requests were coupled with “active choice” being offered by staff (e.g., resident rings call light and staff enter and ask, “What do you need?”). Staff honored participants’ requests for care, when expressed either spontaneously or in response to staff prompts, between 87% and 93% of the time across the four care areas. However, participants seldom reacted to “no choice offered” with noncompliance or otherwise expressing an alternative preference (range 0%–4% across the four care areas). Instead, 80%–93% of participant responses across the four areas were coded as “no conversation at all” or as “general conversation unrelated to care,” which suggests passive compliance in the context of no choice being offered by staff.

Discussion

This research extended a previous study to develop a standardized observational protocol to assess staff–resident interactions pertaining to choice during care provision (Schnelle et al., 2009a, 2009b). Results demonstrated that the observational tool is reliable when administered by research personnel with approximately 3 hr of training. The standardized tool used in this study is now publicly available (www.VanderbiltCQA.org). We also developed a simplified version of the tool for use by nursing home staff. As shown in Figure 1 (www.VanderbiltCQA.org), this simplified version is less time consuming to complete and, thus, captures less detailed information but still yields a reliable assessment of staff–resident interactions related to choice. This protocol differs from the “research” protocol in that it does not distinguish among the types of choice offered (active vs. passive) and is not designed to capture the resident’s response. However, it does assess the central issue: Did the staff member offer the resident choice in each of the four morning care areas? Had this protocol been used to assess morning care in the two nursing homes that participated in this study, the key result would have been the same: The majority of observations (70%) would have been coded as “no choice.” Specifically, the interrater agreement between trained research and supervisory nursing home staff with the abbreviated tool ranged from 86% (toileting) to 100% (dining location) across the four care activities for presence/absence of choice. A research nurse accompanied a supervisory staff member during one morning hour and focused observations on a group of five to six residents on the same hallway to make supervisory observations more time efficient. When a nurse aide entered a resident’s room to provide care, the observation time was less than 10 min per person, although the full hour was needed to identify when staff entered the room.

Figure 1.
Supervisor observational tool: offering choice to residents.

We recommend that nursing home supervisors use this tool to assess whether direct care staff routinely offer choice to residents during morning care provision. If areas for improvement are noted, this assessment can help inform quality improvement efforts. Instructions and tips for conducting observations using the tool are presented in Figure 2. This assessment tool can be useful for nursing homes that have policies stating that residents should be offered choices but lack data independent of staff self-report that choice is routinely being provided at the point of care. Additionally, although the tool featured in this study focuses on four aspects of morning care, the same assessment principles could be applied to assessing choice in other daily care areas (e.g., which social activities to attend, what time and with whom to dine, when to go to bed at night). However, it should be noted that choices about daily care activities represent only one component of resident-centered care. Culture change models more broadly emphasize a home-like environment, privacy, autonomy, and personal freedom (Bowers et al., 2007).

Figure 2.
Instructions for the supervisor observation form.

That there is a need for such assessments is suggested by the results from this preliminary study. Our findings for a sample of nursing home residents showed that staff infrequently offered residents choices during morning care, even though the participating facilities subscribed to the importance of culture change and had conducted prior staff training on offering choice. An important study limitation is the small sample in only two sites. However, findings related to lack of choice during most morning care episodes are consistent with the results of a previous study conducted in 20 facilities in five states, which also showed that staff rarely offered choice during daily care (Schnelle et al., 2009a, 2009b). Again, however, the need to conduct assessment observations should be driven by a desire to determine whether staff within a particular facility offer and honor residents’ choices, for this is an important first step toward ensuring that they do.

Other evaluation findings shed light on current practice in this area. Despite infrequent staff offers of choice, the observational data showed that residents rarely made spontaneous requests and, in most cases, passively complied with staff care practices. While it is possible that some residents do not want staff to offer daily choices, participants’ care plans did not show an assessment of their preferences in any of the targeted care areas. Thus, it seems unlikely that the care routines were determined by resident preference assessments. Results also showed that nursing home staff almost always responded positively and promptly when a resident expressed a clear preference for care. However, such requests occurred infrequently, and in their absence, staff should actively elicit preferences if the goal is to provide resident-directed care. It could be suggested that staff already were aware of residents’ preferences such that it was unnecessary to ask. However, the definition of active choice communications applied in this study was as simple as the staff asking the resident if they were ready to get out of bed or needed to use the bathroom prior to providing the care. We believe this type of communication style during care provision not only encourages choice but also enhances autonomy and dignity. Given that neither site had routine staff–resident assignments, it is unlikely that staff were so knowledgeable of the residents’ preferences that asking was unnecessary. In summary, the observational protocol described in this study provides a reliable assessment tool necessary to empirically support quality improvement efforts toward resident-directed care. The use of a standardized tool will also allow meaningful evaluations of culture change initiatives in the long-term care setting for at least one central aspect of care—allowing residents choices during daily care provision.

Funding

This research was supported by the National Institute of Health (RO1AG032446).

References

  • Bowers B, Nolet K, Roberts T, Esmond S. Implementing change in long-term care: A practical guide to transformation. New York: Commonwealth Fund; 2007. Retrieved from http://www.commonwealthfund.org/Content/Innovations/Tools/2009/Apr/Implementing-Change-in-Long-Term-Care-Bowers.aspx.
  • Centers for Medicare and Medicaid Services. Nursing homes—Issuance of revisions to interpretive guidance at several tags, as part of appendix PP, state operations manual (SOM), and training materials. Baltimore: Author; 2009. Letter to State Survey Agency Directors. Retrieved from http://www.cms.gov/SurveyCertificationGenInfo/downloads/SCletter09_31.pdf.
  • Centers for Medicare and Medicaid Services. Minimum data set, version 3.0 (MDS 3.0) Washington, DC: Author; 2010. Retrieved February 10, 2011, from https://www.cms.gov/NursingHomeQualityInits/45_NHQIMDS30TrainingMaterials.asp#TopOPage.
  • Deming WE. Out of the crisis. Cambridge: Massachusetts Institute of Technology, Center for Advanced Engineering Study; 1986.
  • Doty MM, Koren MJ, Sturla EL. Culture change in nursing homes: How far have we come? Findings from The Commonwealth Fund 2007 National Survey of Nursing Homes. New York: Commonwealth Fund; 2008. Retrieved from http://www.commonwealthfund.org/Content/Publications/Fund-Reports/2008/May/Culture-Change-in-Nursing-Homes–How-Far-Have-We-Come–Findings-From-The-Commonwealth-Fund-2007-Nati.aspx.
  • Molloy DW, Alemayehu E, Roberts R. Reliability of a standardized mini-mental state examination compared with the traditional mini-mental state examination. American Journal of Psychiatry. 1991;148:102–105. [PubMed]
  • Morris JN, Fries BE, Morris SA. Scaling ADLs within the MDS. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences. 1991;54:M546–M553. [PubMed]
  • Neily J, Mills PD, Young-Xu Y, Carney BT, West P, Berger DH, et al. Association between implementation of a medical team training program and surgical mortality. Journal of the American Medical Association. 2010;304:1693–1700. [PubMed]
  • Pronovost P, Needham D, Berenholtz S, Sinopoli D, Chu H, Cosgrove S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. New England Journal of Medicine. 2006;355:2725–2732. Erratum in: New England Journal of Medicine, 2007; 21:356(25):2660. [PubMed]
  • Pronovost P, Vohr E. Safe patients, smart hospitals: How one doctor’s checklist can help us change health care from the inside out. New York: Hudson Street; 2010.
  • Quality Partners of Rhode Island. Care practices. 2011. Retrieved February 17, 2011, from http://www.qualitypartnersri.org/cfmodules/objmgr.cfm?Obj=NHQ_PublicReporting&pmid=124&mid=142&cid=142.
  • Ryan RM, Deci EL. Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist. 2000;55:68–78. [PubMed]
  • Schnelle JF, Bertrand R, Hurd D, White A, Squires D, Feuerberg M, et al. The importance of standardized observations to evaluate nutritional care quality in the survey process. Journal of American Medical Director Association. 2009a;10:568–574. [PubMed]
  • Schnelle JF, Bertrand R, Hurd D, White A, Squires D, Feuerberg M, et al. Resident choice and the survey process: The need for standardized observation and transparency. The Gerontologist. 2009b;49:517–524. [PubMed]

Articles from The Gerontologist are provided here courtesy of Oxford University Press