Efforts to improve nursing home quality have increasingly focused on shifting from task-focused to person-centered care (PCC). PCC incorporates resident preferences and customary routines into care (Elliot, Cohen, Reed, Nolet, & Zimmerman, 2014; Koren, 2010) and has been associated with positive outcomes, such as improved quality of life and reduced behaviors in persons with dementia in some studies (Shier, Khodyakov, Cohen, Zimmerman, & Saliba, 2014). A central tenet of PCC maintains residents are autonomous people capable of making choices and directing care (Lehning & Austin, 2010; McCormack, Roberts, Meyer, Morgan, & Boscart, 2012). PCC facilities have opportunities for choice structured into the care environment. For example, practices in the areas of dining, bathing, sleep/wake routines, and timing of medication administration are modified to provide increased choice and flexibility to support resident choice (Sturdevant, Mueller, & Buckwalter, 2018).
Understanding resident preferences for daily care and activities is fundamental to delivering PCC. Two assessments used to understand resident preferences are the Preferences for Everyday Living Inventory (PELI) (Curyto, Van Haitsma, & Towsley, 2015; Preference Based Living, 2019) and the Minimum Data Set (MDS) 3.0 preference assessment tool (Centers for Medicare & Medicaid Services [CMS], 2018; Housen et al., 2009). Both assessments assess the importance of various preferences and having choices and options regarding aspects of daily living. A PCC toolkit, promoted by the National Nursing Home Quality Improvement Campaign (NNHQIC; 2019), includes both a resident preference assessment tool and an assessment of satisfaction with how preferences are met.
Prior research on preference assessment has shown assessments are useful for understanding important dimensions regarding individual resident preferences. Studies have demonstrated which resident characteristics predict individual preferences (Roberts et al., 2018), how individuals' preferences change over time (Abbott, Heid, Kleban, Rovine, & Van Haitsma, 2018), and the situational dependencies that influence individuals' preferences (Abbott, Bangerter, Humes, Klumpp, & Van Haitsma, 2017). Although understanding individual preferences is critical to PCC delivery, it is less clear how assessment data can be used to identify patterns in resident preferences to support PCC planning at a group or community-wide level or to inform whether resources are in place to meet the most likely preferences of residents. Knowledge of resident characteristics that predict preference patterns could also be used to target changes or care options to appropriate residents. The current study aims to (a) group residents by important preferences for daily care and activities and (b) determine resident characteristics that predict membership in preference groups.
Data for the current study were drawn from admission MDS assessments collected between October 1, 2011 and December 31, 2011. Residents age ≥65 and able to answer the MDS self-report questions for preferences, cognition, and mood were included. Residents who were comatose, had more than one admission in 2011, or had family or staff report their preferences were excluded. After inclusion and exclusion criteria, the original data set of approximately 2.7 million residents was reduced to a final set of 244,718 residents for the current study (additional detail about sample reduction can be found in Roberts et al. ). Characteristics of the final sample are provided in Table 1.
Sample Characteristics (N = 244,718)
Preferences. The MDS preference assessment was used, which assesses the importance of 16 daily care and activity preferences using a 5-point Likert scale with responses 1 to 4 ranging from very important to not important and the fifth response option as important but can't do or no choice (Housen et al., 2009).
Function. Function was determined using the MDS ADL-Long Form summary score, which is calculated based on required level of assistance as assessed by staff on seven activities of daily living reported in MDS Section G. Scores range from 0 to 28, with higher scores indicating more impairment (Morris, Fries, & Morris, 1999; Wysocki, Thomas, & Mor, 2015).
Depression. Probable depression was determined using the total score on the Patient Health Questionnaire-9 (PHQ-9) from MDS Section D. The PHQ-9 screens for signs and symptoms of depression using the presence and frequency of seven mood symptoms. For this analysis, scores were dichotomized as depressed (scores ≥10) or not depressed (scores ≤9) (Saliba, DiFilippo, et al., 2012).
Cognitive Impairment. Cognitive impairment was determined using the total severity score on the Brief Interview for Mental Status (BIMS) in MDS Section C. BIMS assesses recall and temporal orientation with three questions and scores can range from 0 to 15. For this study, scores were trichotomized as no cognitive impairment (scores 13 to 15), moderate impairment (scores 8 to 12), and severe impairment (scores 0 to 7) (Saliba, Buchanan, et al., 2012).
Sociodemographics. Age, race, ethnicity, gender, and marital status were included in the current study.
A latent class model was estimated using Mplus 7.1 (Muthén & Muthén, 2011). Latent class analysis classifies individuals into subgroups based on their response patterns. Preference response options were dichotomized for analysis. Responses of very important, somewhat important, and important but can't do or no choice were collapsed into important. Responses of not very important and not important at all were collapsed into unimportant. The Akaike Information Criterion, Bayesian Information Criterion, Lo-Mendel Rubin test, entropy, interpretability, and class sizes ≥10% were criteria that guided model selection. After the best fitting model was identified, resident characteristics including depression, cognitive impairment, function, and sociodemographics were added as predictors of latent class membership.
A two, three, four, five, and six latent class solution was tested. After the six-class solution, the model could not be reliably estimated due to extreme values. Fit statistics suggested the model fit was improved by the addition of each class, but entropy decreased (Table 2). A four-class solution was determined to be the best fit as it was both interpretable and had class sizes ≥10%.
Latent Class Fit Statistics
The first class comprised 38.3% of the sample and was characterized by very high probability of endorsing all care and activity preferences as important. This class was labeled important. The second and third classes comprised 27.1% and 24.1% of the sample, respectively. Comparing response patterns between these two classes, individuals in the second class were more likely to endorse activity preferences as important and individuals in the third class were more likely to endorse daily care preferences. Class two was labeled activity and class three was care. Finally, the fourth class comprised 10.4% of the sample. The probability of individuals endorsing preferences as important in this group was generally not high and residents were likely to endorse more than one half of preferences unimportant. This group was labeled unimportant. (Figure 1).
Item response probabilities across classes for preference rating of important.
Some resident characteristics significantly predicted class membership. Odds ratios (OR) that were significant and at least a small effect size (OR ≥1.44 and OR ≤0.70) are shown in Figure 2. Cognitive impairment and depression were factors that predicted membership in the unimportant class relative to other classes. African American race and Hispanic ethnicity were predictors of membership in the important class relative to most other classes. Residents in the important class were more likely to have cognitive impairment than residents in the care class. Residents in the activity class were more likely to have cognitive impairment than residents in the care class.
Significant odds ratios comparing resident characteristics between classes.
The current study suggests classifying residents into groups according to preferences is possible and could inform group or community-wide efforts to meet resident preferences or target care options for certain types of residents. A four-class solution was chosen as the best fit and included an important, unimportant, care, and activity class. Findings show some resident characteristics, particularly race, ethnicity, cognition, and depression, may be useful for predicting which preference group an individual most likely belongs.
Most items in the MDS preference assessment evaluate the importance of making daily care choices and having activity options. Residents in the important class, who are likely to endorse all items as important, may be individuals who want to make choices and direct their own day-to-day living. Therefore, a distinctive pattern of important responses across items may be an indicator of a strong desire for choice. The all important response pattern may also suggest these residents have a strong preference for choice in other aspects of their daily living (e.g., preference for timing, method of medication administration) that may be important to assess and consider when organizing work to meet resident preferences. Consistent with the central tenets of PCC, this group of residents may benefit from individualized attention to preferences. Overall, providing environmental support and facilitating a wide variety of choices may be beneficial to these residents.
Race and ethnicity predicted membership in the important class relative to others. African American and Hispanic individuals had greater odds than Caucasian individuals of belonging to the important class relative to other classes. Research has shown a disparity exists between these groups regarding major health care decisions and preferences. For example, studies on advance directives have shown African American and Hispanic individuals are less likely to be asked their preferences for life-sustaining treatment (Degenholtz, Arnold, Meisel, & Lave, 2002; Hopp & Duffy, 2000) and they are more likely than Caucasian individuals to want advanced treatment at end of life (Barnato, Anthony, Skinner, Gallagher, & Fisher, 2009; Hopp & Duffy, 2000). Disparities in other major health care preferences may underscore the need for thorough assessment of daily care and activity preferences when caring for individuals from underrepresented groups. However, more research is needed to better understand the relationship of race/ethnicity, daily care, and activity preferences.
In contrast to the important class, the unimportant class may comprise residents who have a disinterest in making choices and having options about daily care and activities. The factors that predicted membership in this group further suggest this group of residents may be apathetic or disengaged. Residents in this group had higher odds of having cognitive impairment and being depressed, two conditions for which apathy may be a related symptom. It is feasible that residents with cognitive impairment or dementia who experience apathy may have little interest in making choices or directing care (Ellis, Doyle, & Selvarajah, 2016). It is also possible that residents with cognitive impairment have fundamentally different preferences than other residents. However, more research is needed to determine the impact of cognitive impairment on daily preferences. It is also possible that residents with cognitive impairment have difficulty responding to preference questions. Instructions for the preference assessment indicate any resident who is not considered rarely or never understood should be approached to complete the assessment (CMS, 2018). Some residents with severe cognitive impairment may still be understood but may not fully understand the response scale. Proxy responders can underestimate some resident preferences (Heid, Bangerter, Abbott, & Van Haitsma, 2017) and asking residents to report their own preferences is preferred. However, it may be useful for staff members to triangulate resident preference assessment data with additional information from family members or their own observations, as they get to know residents, to gain a rich understanding of preferences.
It is not an unexpected finding that residents with depression are more likely to be in the unimportant class. The depression assessment asks if residents have had less interest in doing things, have little energy, or have had trouble concentrating. These symptoms may also lead to less interest in daily care and activities. Evaluating and treating depression is critical for these residents. Preference assessments are only required during comprehensive MDS assessments (i.e., annually). Reassessing sooner for individuals who have depression may be valuable.
Finally, residents in the unimportant group may be less likely to endorse preferences as important, which may reflect a true lack of preference for choice regarding daily care and activities. This finding could have practical implications for residents in highly PCC environments if they are provided choices and options they do not want to make or have. Further research is needed to better understand whether this pattern reflects a true lack of preference for choice and the impact of PCC environments on outcomes for these residents.
Preference patterns in the care and activity classes have direct relevance for informing PCC practices. It may be difficult to discern which of these groups a resident clearly belongs to because there is a high probability of endorsing most preferences important for both groups. Considering the most remarkable differences as a guide, one may suspect residents belong to the care group if they mark religious activities, being around animals, and being in groups as unimportant. Residents in the activity group may mark using the phone in private and locking belongings as unimportant.
Cognitive status was a significant predictor of membership in the care and activity groups. Residents who belonged to the activity class had higher odds of having cognitive impairment than residents in the care class. One possible explanation for this may be expected length of stay and the types of residents who are likely to have long or short stays. Long-stay residents may be more interested in activities as a source of meaningful engagement and long-stay residents may be more likely to have cognitive impairment. Short-stay residents getting physical rehabilitation may prioritize daily care and be less interested in activities due to rigorous therapy schedules and less need for activity engagement in the short term. However, it was outside the scope of the current study to include expected length of stay, and additional research is needed to explore the reasons some residents are more likely to endorse care or activity preferences.
There are several limitations to the current study. First, only admission assessments were included. The latent class structure may look different if data from other points in time were used. Future studies examining data at other points in a resident's stay are needed. Second, only residents age ≥65 were included. Results may not be generalizable to younger residents. Finally, the entropy value of the model was somewhat low, indicating there was some overlap or poor separation of classes. This finding suggests the individual preference assessment is essential and larger grouping may only be useful as a rough guide for planning preferences at a group level. More research examining whether the number and types of preference responses are relevant to a clinically important outcome or factor (e.g., quality of life, psychosocial well-being) may further inform the practical significance of the classes.
Implications for Nurses
A major implication of the current study is the identification of groups of residents who may need enhanced preference assessments. Conducting more in-depth assessments, such as the PELI, or making a deliberate effort to seek out and learn the specific preferences of residents from underrepresented groups may be needed for nurses to identify unique PCC needs. Using systematic approaches, particularly on dementia units, of soliciting supporting information from families and/or conducting observations of behavior of residents with cognitive impairment may also be an important way to identify the preferences of these residents. The association of depression and preferences highlights an opportunity for nurses and social workers to collaborate on assessment and prompt treatment of depression. Once treated, repeat preference assessments may identify new information for care planning not previously provided. It may also be helpful for nursing staff to provide additional encouragement or support for these residents to engage in daily care or activities.
Results also highlight opportunities for directed communication among nursing staff about resident preferences and suggest ways to organize tasks to support preferences. Residents in the important group may have a strong overall desire to have choices and options. Communicating this desire and indicating known preferences systematically and directly among staff by noting it on nurse aide flowsheets or care cards may help ensure the needs of this group are positively supported. Finally, identifying whether residents belong to the care or activity group can help with organizing nursing tasks and care to meet the different needs of each group. Supporting residents in these groups could be done by working with activity staff to support activity involvement for residents in the activity group (e.g., by scheduling nursing or care tasks around activities whenever possible).
The four groups found in the current study show residents cluster according to overall interest or disinterest in having choices about daily care and activities or specific interest in either care or activities. Recognizing these groups and understanding the types of residents who belong to them can be useful information for guiding approaches to anticipating and planning for resident preferences and needs.
- Abbott, K.M., Bangerter, L.R., Humes, S., Klumpp, R. & Van Haitsma, K. (2017). “It's important, but…”: Perceived barriers and situational dependencies to social contact preferences of nursing home residents. The Gerontologist, 58, 1126–1135. doi:10.1093/geront/gnx109 [CrossRef]
- Abbott, K.M., Heid, A.R., Kleban, M., Rovine, M.J. & Van Haitsma, K. (2018). The change in nursing home residents' preferences over time. Journal of the American Medical Directors Association, 19, 1092–1098. doi:10.1016/j.jamda.2018.08.004 [CrossRef]
- Barnato, A.E., Anthony, D.L., Skinner, J., Gallagher, P.M. & Fisher, E.S. (2009). Racial and ethnic differences in preferences for end-of-life treatment. Journal of General Internal Medicine, 24, 695–701. doi:10.1007/s11606-009-0952-6 [CrossRef]
- Centers for Medicare & Medicaid Services. (2018). Long-term care facility resident assessment instrument 3.0 user's manual. Version 1.16. Retrieved from https://downloads.cms.gov/files/1-MDS-30-RAI-Manual-v1-16-October-1-2018.pdf
- Curyto, K., Van Haitsma, K.S. & Towsley, G.L. (2016). Cognitive interviewing: Revising the preferences for everyday living inventory for use in the nursing home. Research in Gerontological Nursing, 9, 24–34. doi:10.3928/19404921-20150522-04 [CrossRef]
- Degenholtz, H.B., Arnold, R.A., Meisel, A. & Lave, J.R. (2002). Persistence of racial disparities in advance care plan documents among nursing home residents. Journal of the American Geriatrics Society, 50, 378–381. doi:10.1046/j.1532-5415.2002.50073.x [CrossRef]
- Elliot, A., Cohen, L.W., Reed, D., Nolet, K. & Zimmerman, S. (2014). A “recipe” for culture change? Findings from the THRIVE survey of culture change adopters. The Gerontologist, 54(Suppl. 1), S17–S24. doi:. doi:10.1093/geront/gnt133 [CrossRef]
- Ellis, J.M., Doyle, C.J. & Selvarajah, S. (2016). The relationship between apathy and participation in therapeutic activities in nursing home residents with dementia: Evidence for an association and directions for further research. Dementia, 15, 494–509. doi:10.1177/1471301214527300 [CrossRef]
- Heid, A.R., Bangerter, L.R., Abbott, K.M. & Van Haitsma, K. (2017). Do family proxies get it right? Concordance in reports of nursing home residents' everyday preferences. Journal of Applied Gerontology, 36, 667–691. doi:10.1177/0733464815581485 [CrossRef]
- Hopp, F.P. & Duffy, S.A. (2000). Racial variations in end-of-life care. Journal of the American Geriatrics Society, 48, 658–663. doi:10.1111/j.1532-5415.2000.tb04724.x [CrossRef]
- Housen, P., Shannon, G.R., Simon, B., Edelen, M.O., Cadogan, M.P., Jones, M. & Saliba, D. (2009). Why not just ask the resident? Refinement of a preference assessment tool for nursing homes. Journal of Gerontological Nursing, 35(11), 40–49. doi:10.3928/00989134-20091001-01 [CrossRef]
- Koren, M.J. (2010). Person-centered care for nursing home residents: The culture-change movement. Health Affairs, 29, 312–317. doi:10.1377/hlthaff.2009.0966 [CrossRef]
- Lehning, A.J. & Austin, M.J. (2010). Long-term care in the United States: Policy themes and promising practices. Journal of Gerontological Social Work, 53, 43–63. doi:10.1080/01634370903361979 [CrossRef]
- McCormack, B., Roberts, T., Meyer, J., Morgan, D. & Boscart, V. (2012). Appreciating the ‘person’ in long-term care. International Journal of Older People Nursing, 7, 284–294. doi:10.1111/j.1748-3743.2012.00342.x [CrossRef]
- Morris, J.N., Fries, B.E. & Morris, S.A. (1999). Scaling ADLs within the MDS. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 54, M546–M553. doi:10.1093/gerona/54.11.M546 [CrossRef]
- Muthén, L.K. & Muthén, B.O. (2011). Mplus user's guide (6th ed.). Los Angeles, CA: Authors.
- National Nursing Home Quality Improvement Campaign. (2019). Person-centered care. Retrieved from https://www.nhqualitycampaign.org/goalDetail.aspx?g=PCC#tab2
- Pioneer Network. (2018). What is culture change. Retrieved from https://www.pioneernetwork.net/elders-families/what-is-culture-change/
- Preference Based Living. (2019). Tools. Retrieved from https://preferencebasedliving.com/?q=peli-tools
- Roberts, T.J., Gilmore-Bykovskyi, A., Lor, M., Liebzeit, D., Crnich, C.J. & Saliba, D. (2018). Important care and activity preferences in a nationally representative sample of nursing home residents. Journal of the American Medical Directors Association, 19, 25–32. doi:10.1016/j.jamda.2017.06.028 [CrossRef]
- Saliba, D., Buchanan, J., Edelen, M.O., Streim, J., Ouslander, J., Berlowitz, D. & Chodosh, J. (2012). MDS 3.0: Brief interview for mental status. Journal of the American Medical Directors Association, 13, 611–617. doi:10.1016/j.jamda.2012.06.004 [CrossRef]
- Saliba, D., DiFilippo, S., Edelen, M.O., Kroenke, K., Buchanan, J. & Streim, J. (2012). Testing the PHQ-9 interview and observational versions (PHQ-9 OV) for MDS 3.0. Journal of the American Medical Directors Association, 13, 618–625. doi:10.1016/j.jamda.2012.06.003 [CrossRef]
- Shier, V., Khodyakov, D., Cohen, L.W., Zimmerman, S. & Saliba, D. (2014). What does the evidence really say about culture change in nursing homes?The Gerontologist, 54(Suppl. 1), S6–S16. doi:10.1093/geront/gnt147 [CrossRef]
- Sturdevant, D.L., Mueller, C.A. & Buckwalter, K.C. (2018). Measurement of nursing home culture change: Systematic review. Research in Gerontological Nursing, 11, 103–112. doi:10.3928/19404921-20171205-01 [CrossRef]
- Wysocki, A., Thomas, K.S. & Mor, V. (2015). Functional improvement among short-stay nursing home residents in the MDS 3.0. Journal of the American Medical Directors Association, 16, 470–474. doi:10.1016/j.jamda.2014.11.018 [CrossRef]
Sample Characteristics (N = 244,718)
|Age (mean, SD)||81.02 (8.2)|
|Female (n, %)||160,053 (65.4)|
|Race/ethnicity (n, %)|
| Caucasian||207,767 (87.1)|
| African American||18,833 (7.9)|
| Hispanic||8,146 (3.4)|
| Othera||3,915 (1.6)|
|Married (n, %)||80,770 (33.6)|
|Depressed (n, %)||17,361 (7.1)|
|Cognitive impairment (n, %)|
| None||153,540 (62.9)|
| Moderate||57,274 (23.5)|
| Severe||33,253 (13.6)|
|Functionb (mean, SD)||16.63 (4.6)|
Latent Class Fit Statistics