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.
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
). 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).