Untangling the complex web of avoidable nursing home-to-hospital transfers of residents with dementia
Abstract
INTRODUCTION
Nursing home (NH) residents with Alzheimer's disease or related dementias (ADRD) are at high risk for hospital transfer. We aimed to (1) describe characteristics and predictors of avoidable transfer of residents with ADRD and (2) explore how “what matters” influences the decision to transfer.
METHODS
We applied an exploratory, mixed methods design using data collected as part of a Centers for Medicare and Medicaid Services demonstration project. Advanced practice registered nurses documented retrospective details about nursing home (NH)-to-hospital transfers (n = 3687) from 16 NHs.
RESULTS
NH residents with ADRD had 1.22 times higher odds of having an avoidable NH-to-hospital transfer (odds ratio = 1.22; 95% confidence interval = 1.03, 1.45). Factors contributing to avoidable transfers were age, stage of ADRD, what matters to the resident and their family, changes in condition, and resources available in the NH.
DISCUSSION
These findings highlight the need for enhanced specificity in the discussion and documentation of resident and family preferences and continued investments in the NH workforce.
Highlights
- This article reports on factors contributing to avoidable nursing home (NH)-to-hospital transfer of residents with Alzheimer's disease and related dementias (ADRD). The mixed methods design used in this study offers insight beyond what is possible using a single-method design.
- Using data collected from a Centers for Medicare and Medicaid Services demonstration project, advanced practice registered nurses documented retrospective details about NH-to-hospital transfers (n = 3687) of residents.
- NH residents with ADRD were more likely to have an avoidable NH-to-hospital transfer. Factors contributing to avoidable transfers were age, stage of ADRD, what matters to the resident and their family, changes in condition, and resources available in the NH.
- These findings highlight the need for enhanced specificity in discussion and documentation of resident and family preferences and continued investments in the NH workforce.
1 INTRODUCTION
Close to two thirds of the > 1.3 million nursing home (NH) residents in the United States are cognitively impaired, often with Alzheimer's disease or related dementias (ADRD)1 and are at high risk of hospital transfer.2, 3 Reducing avoidable hospitalizations for NH residents is a national priority due to the negative effects on resident health and high costs to Medicare and Medicaid.4-6 NH residents with ADRD are at increased risk for falls,7 delirium,8 functional decline,9 malnutrition, and dehydration10 which may lead to injury11 and increased risk of hospitalization.12-15 NH-to-hospital transfers result in expenditures of > $2.6 billion annually,16 harm to residents from the physical and emotional stress of relocation,17, 18 and are avoidable in as many as 60% of cases.19-21 Avoidable NH-to-hospital transfers of NH residents refer to acute conditions that could have been prevented with appropriate care or conditions that could be effectively managed in the NH.19, 22 Avoidable hospitalizations disrupt plans of care, negatively impact quality of life, and can lead to physical and cognitive decline.6, 23
Empirical evidence on factors contributing to avoidable NH-to-hospital transfers of residents with ADRD is sparse and mixed. Givens et al.2 evaluated potentially avoidable hospitalizations of NH residents with advanced dementia from 22 Boston area hospitals and found suspected infections, feeding tube complications, and lack of do not resuscitate (DNR) orders were significantly associated with hospital transfer. Nakashima et al.23 examined the impact of do not hospitalize (DNH) orders on resident transfers and found residents with dementia and DNH orders had significantly fewer hospital stays but not emergency department visits. Notwithstanding, a 2019 study using national Centers for Medicare and Medicaid Services (CMS) data found residents with ADRD had a lower relative risk of hospitalization, both avoidable and non-avoidable, compared to residents without ADRD; however, in considering whether a hospitalization was avoidable, the study used only hospital diagnoses thought to represent avoidable conditions and did not account for contextual factors.24 Multiple studies have concluded that opportunities remain to further reduce avoidable transfers in this population and improve goal-directed care.23, 25, 26 We found few, if any, studies focused on predictors of avoidable NH-to-hospital transfer of residents that examined contextual factors including how resident and family preferences impact avoidable transfers.
By age 80, ≈ 75% of individuals with ADRD will live in a NH1, making this an important care setting for this vulnerable population. Aligning care of NH residents with ADRD with “what matters” most to them is a key attribute of the Age-Friendly Health System Initiative. The product of a collaboration between the John A. Hartford Foundation and the Institute for Healthcare Improvement, key attributes of the Age-Friendly Health System Initiative include valuing patient goals and preferences, supporting family caregivers and including them in the treatment plan, effective and efficient communication, interaction among and across settings and caregivers, and optimizing value.27 Meeting these standards will likely enhance the quality of care for older adults with ADRD.
Our study builds knowledge in this area by leveraging data from a 5-year CMS demonstration project aimed at reducing avoidable hospital transfer of NH residents and using qualitative methods to help enhance the meaning/clinical relevance of quantitative results. The purpose of this study was to (1) describe characteristics and predictors of avoidable NH-to-hospital transfer of residents with ADRD and (2) explore how “what matters” influences the decision to transfer.
2 METHODS
2.1 Design
We used a retrospective, exploratory, mixed methods design. Quantitative data were collected and analyzed first; then qualitative data were used to understand and elaborate on the quantitative findings. We integrated quantitative and qualitative results during the analysis and interpretation of study results.
2.2 Data
Data for this analysis were collected over 5 years (2016–2020) as part of the Missouri Quality Initiative (MOQI) CMS Innovations Center demonstration project. The MOQI project aimed to reduce avoidable hospitalizations of NH residents and improve quality of care28. NHs (n = 16) that participated in the MOQI demonstration project had access to similar resources including (1) advanced practice nurses (APRN) working as essential staff, providing clinical expertise to front-line clinicians and support for facility-level change efforts; (2) data feedback reports about hospitalizations and transfers and use of Interventions to Reduce Acute Care Transfers (INTERACT) tools;29 and (3) enhanced communication systems (e.g., secure text messaging) used to facilitate health information exchange. Every time a resident was transferred from the NH to the hospital, regardless of the outcome of the transfer (e.g., emergency department visit only, hospital admission, etc.), APRNs completed a Qualtrics survey based on the INTERACT tool. Broad categories of the survey included resident characteristics, risk factors for hospitalization, acute change in condition, non-clinical factors contributing to the transfer, and action(s) taken to evaluate and manage the change in condition before transfer. The Qualtrics survey contained both closed and open-ended responses. NHs that participated in the project ranged in bed size from 120 to > 300 beds, and were located in metropolitan, micropolitan, and rural settings (see Table 1). Residents enrolled in the MOQI project were dual eligible for Medicare and Medicaid services. Short-stay residents, defined as those residing in the NH < 100 days were excluded from the demonstration project.
Sample NHs (n = 16) | Total US NHs (N = 15,600) | |
---|---|---|
Location | ||
Metropolitan | 10 (62%) | 11,154 (71%) |
Micropolitan | 3 (19%) | 2168 (14%) |
Small town/rural | 3 (19%) | 2277 (15%) |
Ownership | ||
Non-profit | 2 (12%) | 3744 (24%) |
For-profit | 14 (88%) | 11865 (76%) |
Average number of certified beds (mean) | 120–321 (198) | 2–1,389 (106) |
Medicaid-certified | 16 (100%) | 14,820 (95%) |
Medicare-certified | 16 (100%) | 15,288 (98%) |
2.3 Approach to analysis
2.3.1 Quantitative
Determination of the outcome measure (avoidable NH-to-hospital transfer) was made by a team consisting of the APRN, project coordinator, and APRN supervisor. Using an iterative interrogative approach, the team explored the root causes of each transfer until a consensus was met on which transfers were deemed avoidable versus unavoidable20. In the primary study, data were collected only for residents who were transferred to the hospital; therefore, we used unavoidable transfers as the comparison group. The stage of dementia was based on the Alzheimer's Association's definitions of early (mild), middle (moderate), and late (severe)30. For the quantitative analysis, descriptive statistics were used to characterize the sample using frequencies, means, and standard deviations. Next, we used a generalized linear mixed-effect model (GLMM) with a binary logit link function to account for individual residents who had more than one transfer. In addition to resident characteristics (age, sex, race, cardiopulmonary resuscitation [CPR] status, stage of ADRD), we included NH attributes (bed size, location, type of ownership) in the model to account for nested random effects (i.e., variability of participants nested within the NH).
RESEARCH IN CONTEXT
-
Systematic review: Nursing home (NH)-to-hospital transfers result in expenditures of more than $2.6 billion annually, harm to residents from the physical and emotional stress of relocation, and are avoidable in as many as 60% of cases. Empirical evidence on factors contributing to avoidable NH-to-hospital transfer of residents with Alzheimer's disease and related dementias (ADRD) is lacking; thus, more detailed and contextual input is needed. We conducted a mixed methods study to (1) describe characteristics and predictors of avoidable transfer of residents with ADRD and (2) explore how “what matters” influences the decision to transfer.
-
Interpretation: NH residents with ADRD were more likely to have an avoidable NH-to-hospital transfer. Factors contributing to avoidable transfers were age, cardiopulmonary resuscitation (CPR) status, what matters to the resident and their family, changes in condition, and resources available in the NH.
-
Future directions: To move ADRD research forward, additional studies are warranted to (1) establish consensus on a theoretical definition of “avoidable” transfer; (2) help residents and families fully understand the alignment of CPR status, personal preferences, and realistic expectations of NH care; and (3) assess the impact of current and proposed NH staffing legislation on outcomes such as avoidable NH-to-hospital transfer.
2.3.2 Qualitative
To draw our qualitative sample, first we grouped transfers based on stage of ADRD (early, middle, late) and CPR status (DNR, full code, other). Next, we randomly selected 15 transfers from each group to maximize variation and increase the likelihood that the findings would reflect different perspectives31. Based on our relatively homogenous sample and focused study objectives, we estimated (a priori) needing a sample size of 90 avoidable transfers (15 from each group) to reach saturation, that is, the point in data collection at which all relevant conceptual categories have been identified, explored, and exhausted32.
We analyzed responses to open-ended questions from the Qualtrics survey completed by the APRN after each transfer occurred. Respondents were asked to describe changes in condition or other non-clinical factors that contributed to the hospital transfer. We used directed content analysis33 and a priori categories based on the Age-Friendly Health System Initiative's 4Ms (what matters, medications, mobility, mentation) conceptual framework to begin the analysis. Dedoose qualitative software (https://www.dedoose.com/) was used to organize and visualize data. For each transfer, data were coded by reviewing the survey results line by line. Initially, text from 10 residents were coded jointly by the team to establish consistency. This process involved members of the research team reading through the transfer data together and assigning representative codes. The remaining transfers were coded independently by the primary author (K.P.) and a research assistant (D.P.). Any differences in coding were discussed among the research team until a consensus was met. Once initial codes were assigned, they were reviewed and organized into subcategories under each of the four a priori categories. During the next phase of the analysis, subcategories were merged when overlap was present and new categories were developed when the codes did not fit within one of the a priori categories. The findings were reviewed and agreed upon by the team as presented in section 3.2.
2.3.3 Data management and ethics
Data are stored and handled as recommended by the institutional review board at the University of Missouri. Due to ongoing analyses, the authors cannot share the raw data used in this study. Inquiries about collaborations to use these data can be sent to the corresponding author. All study procedures were approved by the institutional review board at the University of Missouri, Columbia (project number 2092039).
3 RESULTS
3.1 Quantitative
Over the 5-year study period, there were a total of 3687 transfers of which 1818 (49.3%) had a diagnosis of ADRD. Among NH residents with ADRD that transferred, the majority were White (69.3%), female (68.6%), had middle-stage dementia (48.6%), had CPR status of DNR (51.9%), and were potentially avoidable (54.6%; see Figure 1). NH residents with ADRD had 1.22 times higher odds of having an avoidable NH-to-hospital transfer compared to NH residents with other diagnoses, holding other predictors (age, race, sex, CPR status, bed size, location, ownership type) at a fixed value (odds ratio [OR] = 1.22, 95% confidence interval [CI] = 1.03, 1.45). When we included four levels of exposure (no ADRD, early, middle, and late), we found statistically significant predictors of avoidable transfer were age (OR = 1.48, 95% CI = 1.10, 1.99) and late stage ADRD (OR = 1.45, 95% CI = 1.11, 1.88; see Table 2).

Overall n = 3687 | Without ADRD n = 1869 (50.7%) n (%) | With ADRD n = 1818 (49.3%) n (%) | p value | Odds ratio (95% CI) | |
---|---|---|---|---|---|
Age | |||||
< = 64 | 783 | 614 (32.9) | 169 (9.3) | N/A | Reference |
65–74 | 897 | 550 (29.4) | 347 (19.1) | 0.358 | 1.13 (0.87, 1.47) |
75–84 | 959 | 394 (21.1) | 565 (31.1) | 0.142 | 1.23 (0.93, 1.63) |
> = 85 | 1048 | 311 (16.6) | 737 (40.5) | 0.010** | 1.48 (1.10, 1.99) |
Race | |||||
Black or African American | 939 | 423 (22.6) | 516 (28.4) | N/A | Reference |
White | 2678 | 1419 (75.9) | 1259 (69.3) | 0.115 | 0.81 (0.62, 1.05) |
Other | 70 | 27 (1.4) | 43 (2.4) | 0.857 | 1.06 (0.57, 1.96) |
Sex | |||||
Female | 2329 | 1082 (57.9) | 1247 (68.6) | N/A | Reference |
Male | 1358 | 787 (42.1) | 571 (31.4) | 0.563 | 1.05 (0.88, 1.26) |
CPR status | |||||
DNR | 1597 | 659 (35.5) | 938 (51.9) | N/A | Reference |
Full code | 1879 | 1110 (59.8) | 769 (42.5) | 0.077 | 0.84 (0.68, 1.02) |
Other | 190 | 88 (4.7) | 102 (5.6) | 0.482 | 1.17 (0.76, 1.80) |
Stage of dementia | |||||
No dementia | 1869 | 1869 (100) | N/A | N/A | Reference |
Early | 167 | N/A | 167 (13.4) | 0.845 | 0.96 (0.66, 1.40) |
Middle | 606 | N/A | 606 (48.6) | 0.138 | 1.19 (0.95, 1.49) |
Late | 475 | N/A | 475 (38.1) | 0.006** | 1.45 (1.11, 1.88) |
Bed size | |||||
Small (< 160) | 1266 | 693 (37.1) | 573 (31.5) | N/A | Reference |
Medium (160–220) | 797 | 498 (26.6) | 299 (16.4) | 0.718 | 1.12 (0.61, 2.06) |
Large (> 220) | 1624 | 678 (36.3) | 946 (52.0) | 0.344 | 1.35 (0.72, 2.52) |
Location | |||||
Metro | 2684 | 1272 (68.1) | 1412 (77.7) | N/A | Reference |
Micro | 804 | 462 (24.7) | 342 (18.8) | 0.584 | 1.19 (0.64, 2.23) |
Small town/rural | 199 | 135 (7.2) | 64 (3.5) | 0.097 | 0.49 (0.21, 1.14) |
Ownership | |||||
Not for Profit | 243 | 88 (4.7) | 155 (8.5) | N/A | Reference |
For profit | 3444 | 1781 (95.3) | 1663 (91.5) | 0.736 | 0.87 (0.37, 2.00) |
Avoidable transfer | |||||
Yes | 1851 | 863 (46.3) | 988 (54.6) | ||
No | 1822 | 999 (53.7) | 823 (45.4) |
- Abbreviations: ADRD, Alzheimer's disease and related dementias; CI, confidence interval; CPR, cardiopulmonary resuscitation; GLMM, generalized linear mixed-effect model; NH, nursing home.
- ** Significant at alpha 0.05.
3.2 Qualitative
Over the course of the analysis, 29 open codes were developed and organized into three broad themes representing factors contributing to avoidable NH-to-hospital transfer of residents with ADRD: (1) what matters, (2) changes in condition, and (3) resources not available in the NH. Themes, subthemes, and illustrative comments are shown in Table 3. Because stage of ADRD, CPR status, and age were found to be associated with avoidable NH-to-hospital transfers, we included them as descriptors in the qualitative analysis. This approach allowed us to triangulate findings by enriching the quantitative results with contextual data.
Theme | Subtheme | Illustrative quote |
---|---|---|
What matters | Family preferences |
|
Resident preferences |
|
|
Change in condition | Abnormal vital signs |
|
Abnormal lab/diagnostic tests |
|
|
Abnormal physical assessment |
|
|
Mentation | ||
Mobility |
|
|
Other |
|
|
Lack of resources | Physical resources | |
Equipment |
|
|
Lab/diagnostic tests |
|
|
Staff resources | ||
Effective communication |
|
|
NH staff skills/experience |
|
|
Sufficient staff |
|
- Abbreviations: CNA, certified nursing assistant; DNR, do not resuscitate; ER, emergency department; NP, nurse practitioner; UA, urinalysis; WBC, white blood cell.
3.3 What matters
Resident and family preferences were the most common factors influencing the decision to transfer residents with ADRD regardless of resident age, CPR status, and stage of ADRD. Moreover, resident and family preferences were identified more often in residents with CPR status of “other” compared to those with CPR status of full code and DNR. When we explored differences according to family versus resident preferences, we found the majority of avoidable transfers that occurred as the result of the resident's preference were among younger residents (age 57–64).
3.4 Changes in condition
The theme of changes in resident condition includes three subthemes: abnormal vital signs, abnormal lab/diagnostic tests, and abnormal physical assessment. Abnormal vital signs and lab/diagnostic tests were common factors influencing avoidable NH-to-hospital transfers. Deviations from the normal range of temperature, heart rate, and respiratory rate were frequently noted as well as increased pain. Abnormal results from diagnostic blood and urine tests as well as x-ray and scans were frequently cited as contributing factors in avoidable transfers. The subcategory of abnormal physical assessment includes changes in mentation, mobility, and other. Changes in mentation among residents with dementia were common reasons contributing to avoidable transfer, especially for residents with full-code CPR status. APRNs noted changes such as the resident being less conversational, lethargic, and confused. Mobility issues were mostly related to some type of fall, usually occurring during a transfer from bed to chair or wheelchair to bed. The changes in condition subcategory of “other” represents other abnormal physical assessment findings contributing to avoidable transfer such as decreased intake of food/fluids, signs and symptoms of dehydration, and bleeding.
When we consider change in condition in terms of age and CPR status, we found younger residents (age ≤ 64) had more avoidable transfers due to bleeding, mobility, and pain. Among all residents with CPR status of “other” there were more avoidable transfers due to abnormal physical assessment and abnormal lab/diagnostic test results. For residents with CPR status of full code, changes in mentation were documented more frequently as having contributed to the transfer compared to residents with CPR status of other and DNR.
3.5 Resources not available in the NH
Needed resources to manage resident conditions that were not available in the NH include two subcategories: physical resources and staff resources. Lack of access to equipment such as sutures and lab/diagnostic tests such as x-ray and computed tomography were factors contributing to avoidable transfer, especially when the resident experienced a fall. APRNs noted that having access to these types of resources might have prevented the need to transfer the resident altogether. Staff resources include three categories: effective communication, NH staff skills/experience, and sufficient staff. Effective communication refers to situations in which the transfer might have been avoided had there been better communication between NH staff and physician/nurse practitioner or between NH staff and resident family. NH staff skills/experience refers to situations in which transfers could have been avoided if NH staff felt more comfortable/confident in making the decision to transfer or to keep the resident in the facility. The last category of sufficient staff refers to situations in which the resident needed more frequent monitoring than what NH staff could provide. Having a lack of staff resources was noted as a contributing factor to avoidable transfers more often for residents with full-code status. Effective communication, particularly having timely access to advanced directives and hospitalization preferences, was a more common influential factor for the avoidable transfer of younger residents (age ≤ 64).
4 DISCUSSION
This study adds knowledge about avoidable NH-to-hospital transfer of residents with ADRD. We found residents who had a diagnosis of ADRD, were older, and had late-stage dementia were more likely to have an avoidable transfer. When we explored contextual factors contributing to avoidable transfers, we identified three themes: what matters, changes in condition, and resources that were not available in the NH. While other studies have reported similar findings that resident age, code status, and clinical complexity are associated with increased transfers2, 23, 34, few studies have specifically examined avoidable transfers of residents with ADRD and to our knowledge, even fewer have explored contributing factors using the Age-Friendly Health System Initiative 4 M framework27.
4.1 Factors contributing to avoidable NH-to-hospital transfer of residents with ADRD
In our study, NH residents with ADRD were more likely to have an avoidable NH-to-hospital transfer. While few studies have focused on potentially avoidable NH-to-hospital transfers among residents with dementia, there are some we should consider. Temkin-Greener et al. found a lower risk of avoidable NH-to-hospital transfers of residents with ADRD in a national study using CMS data24. Alternatively, Feng et al. found no significant difference in potentially avoidable hospitalizations and ED visits comparing NH residents with and without ADRD35. It is important to note that the determination of “avoidable” transfers differed in these studies compared to ours. In both studies, the determination of avoidable was based on diagnoses thought to represent conditions that could reasonably be managed in the NH. However, this determination does not account for contextual factors. For example, while a diagnosis of urinary tract infection might be considered avoidable in most cases, if the resident's symptoms could not be managed in the NH (e.g., severe pain, gross hematuria), the transfer may have been necessary. The determination of “avoidable” transfers in our study was reached from a consensus of clinical personnel who conducted a root-cause analysis36. Despite the validation of this approach, we recognize limitations related to individual perceptions and interpretation. Given the lack of consensus on a theoretical definition of “avoidable” transfers37, future studies should continue to seek validated measures that maintain flexibility allowing for generalizability to diverse NH settings.
4.2 What matters: resident and family preferences
Our findings that stage of ADRD and age were statistically significantly associated with avoidable NH-to-hospital transfers are better understood considering our qualitative findings. The qualitative analysis revealed that for residents with late-stage dementia, what matters (specifically family preference) and effective communication influenced the transfer decision. While the importance of early and frequent discussions about goals of care and end-of-life preferences are well documented in the literature23, 26, our findings support two specific recommendations: (1) enhanced specificity in discussions of resident and family preferences and (2) improved communication of those preferences.
Associations between CPR status, specifically DNR and DNH orders, among residents with late-stage dementia and avoidable transfers have been reported in the literature. For example, Givens et al. found lack of a DNR order was associated with hospital transfer of residents with advanced dementia2. Nakashima et al. reported NH residents with ADRD who had DNH orders had significantly fewer hospital stays but not emergency department visits23. While our dataset did not contain “DNH” orders, the CPR status “other” implies that the resident's preferences lie somewhere between DNR and full code and could include specific preferences regarding hospitalization. However, making clear exactly what the resident prefers given this CPR status can be challenging. Clarification requires residents and families to fully understand the different CPR statuses, how they align with personal values and preferences, and to have realistic expectations for care provided in the NH. Having discussions about goals of care, especially for older residents and those nearing end of life, have been found to be most successful when an interdisciplinary approach is used38. Confirmation that residents and family fully understand these orders using a multidisciplinary approach (including nurses and social workers when possible) could help ensure resident and family preferences are effectively implemented.
4.3 NH staffing
In our qualitative analysis, we found the skills and experience of NH staff as well as having an adequate number of staff influenced avoidable NH-to-hospital transfers. The decision to transfer a NH resident to the hospital is complex and relies on the detection of subtle changes in condition and timely communication39. Further, the frequency of comorbidities and progressive loss of language that results from ADRD can add complexity surrounding the decision to transfer40. NH staff development could help address this issue including more training on managing chronic conditions and differentiating changes in mentation indicating an acute process rather than the expected progression of ADRD. However, the benefit of enhanced staff development is limited if staffing is inadequate.
For decades, NHs have endured staff shortages and high turnover rates placing additional strain on an insufficient workforce41. While strategic investment in staff and in tools could help to improve outcomes including reduced avoidable hospital transfers, further research is warranted. A recent study found that increasing staffing alone did not improve outcomes at NHs with both low- and high-dementia censuses42. Federal minimum staffing regulations, recently introduced by the Biden administration, aim to address the NH workforce crisis43. This plan involves a $75 million investment to facilitate the recruitment, training, and transition of NH staff. Complementary to existing programs geared toward bolstering the nursing workforce, this initiative might help to alleviate the industry's staffing challenges. However, the regulation's implementation will not be immediate as it is currently undergoing a public response period and, if approved, will be rolled out over a span of 2 to 5 years. Consequently, the extent and timing of its impact on NH staffing and quality of care remain uncertain.
4.4 Limitations
Findings from this study should be interpreted with some key limitations in mind. First, although 16 NHs were included in the study, they are all located in one Midwest state. Because these NHs participated in the MOQI demonstration project, they each had a full-time APRN embedded on site, which also may limit broad generalization. Nonetheless, the NHs in this study faced the same challenges as many NHs when making the decision to transfer residents with ADRD. Second, as a retrospective analysis, the study team could not clarify APRN responses from the Qualtrics survey. Interviews with the APRN completing the survey as well as data from the residents’ medical records could have provided important clarification, for example, specifics of CPR status designation (i.e., what was meant by “other”) and documentation of conversations about goals of care and preferences. Third, because APRNs only documented retrospective details of residents who were transferred to the hospital, we were not able to compare residents who transferred to those who did not. However, we used transfers that were deemed unavoidable to make comparisons between groups. Finally, causality should not be implied as this was an exploratory mixed-methods study.
5 CONCLUSION
In this study, NH residents with ADRD were more likely to experience an avoidable NH-to-hospital transfer. Factors contributing to avoidable transfers were age, stage of ADRD, what matters to the resident and their family, changes in condition, and resources available in the NH. Reducing avoidable NH-to-hospital transfers of residents with ADRD is a priority given the physical and emotional stress it places on residents and associated costs. Avoidable transfers impact the quality of care and quality of life for residents with ADRD. Perhaps more importantly, avoidable transfers are often not aligned with what matters most to residents and their family members. Detailed conversations about preferences including CPR status and hospitalization preferences must take place frequently and be well documented. NHs must offer training and support to ensure staff can recognize situations that can be safely managed in the NH and those requiring transfer. Last, ensuring adequate NH staff is essential for early detection of changes in condition and treatment that could prevent avoidable transfer, ultimately reducing morbidity and mortality in residents with ADRD.
ACKNOWLEDGMENTS
Research reported in this publication was supported by the National Institute on Aging of the National Institutes of Health (award R01AG078281; principal investigator: KRP). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
CONFLICT OF INTEREST STATEMENT
Author disclosures are available in the supporting information.
CONSENT STATEMENT
The study was approved by the institutional review board at the University of Missouri, Columbia, IRB No. 2092039. Participant informed consent was obtained from all subjects and/or their legal representatives via waiver of documentation of informed consent meaning a written signature by each participant was not required. All methods were performed in accordance with the relevant IRB guidelines and regulations.