This study compared the predictive capacity of the new version 4.0 of the NECPAL instrument (8, 9, 15), PROFUND index (17, 18), and PPI (16, 48) at four time points (3, 6, 12, and 24 months) in a sample of elderly nursing home residents with ACC with and without dementia. Higher predictive capacities were found for NECPAL4.0 and PROFUND according to the presence/absence of dementia and measurement time point.
The age and sex profiles of the residents were similar to those in previous studies of this type in the nursing home setting (37, 49–55). The percentage of residents classified with ACD by NECPAL3.1 was somewhat lower than described by Martínez-Muñoz (56) but higher than reported by da Costa et al. (57). The percentage of residents with positive responses to the “surprise” question was higher than in previous studies that included this item in the study of other instruments (54, 55).
Similar proportions of residents with ACC were in NECPAL survival stages I and II and only a small proportion in stage III, implying a lower median survival (3.6 months). A higher proportion of patients were in stage III in the investigation by Calsina-Berna et al. (58), which included a larger percentage of patients receiving palliative care; however, few studies have been published on this issue, which warrants further research.
PROFUND results (7, 18) were in agreement with the findings by Da Costa et al. (57) and Moretti et al. (19) of a high or very high risk of mortality (≥ 7 points) in most residents. With regard to PPI findings, a mean of 4.5 points was obtained by Nieto-Martín et al. (16), indicating a worse survival than predicted in the present series.
There was a lower percentage of deaths in the present study than in some other studies of residents with dementia (60, 61, 23), although these only reported the mortality for periods < 12 months. Studies in nursing homes with a 24-month follow-up by Turrillas et al. (9) and Martínez-Muñoz et al. (56) observed a higher mortality rate (43% and 52.74%, respectively) at 24 months in comparison to the present series (38.4% ). Most residents who died within two years had dementia. Bernabeu-Wittel et al. (17) described the presence of dementia as a predictor of mortality and included it in the PROFUND index. In general, the instruments under study described a worse prognosis for the residents with versus without dementia, although dementia often coexists with other diseases that might affect the prognosis.
In the global sample, the highest AUC was obtained at 3 and 6 months using the PROFUND index (17, 18), with statistical significance, while the highest AUC at 24 months was obtained with NECPAL4.0, with a close-to-significant result, supporting data in the validation study for this instrument (9).
PROFUND obtained higher AUC values in the patients with versus without dementia, reaching statistical significance at 12 and 24 months. This might be attributable to the inclusion of a specific analytical parameter (hemoglobin < 10 g/dL) known to predict a worse prognosis and improve the prognostic accuracy in patients with different diseases, including dementia. NECPAL4.0 showed higher AUC values for the residents without dementia, in line with the finding by the original validation study of a worse predictive capacity in patients with versus without dementia (9). The lowest AUC values were observed with PPI (16) at all follow-up times, and no statistically significant results were observed.
Strengths and limitations
A larger sample of residents would have increased the statistical power, thereby improving the detection of statistical significance in trends identified in the present study. A further limitation was the lack of data on comorbidities in the groups with and without dementia, which should be provided separately in future studies, given that dementia is frequently accompanied by other diseases that can affect the prognosis. It should also be taken into account that part of the survival follow-up coincided with the COVID19 pandemic, although the nursing homes reported that none of the residents in the study died from this cause.
According to these findings, PROFUND is the best instrument to predict survival in nursing home residents with ACC in general and in those with dementia, especially over the short term, whereas NECPAL4.0 offers the best performance in residents without dementia and over the long term (≥ 24 months). Hence, these instruments complement each other in terms of type of resident and time scale. NECPAL4.0 not only covers a longer time period than PROFUND but also offers information on the palliative needs of residents, helping nursing homes to prioritize resources for their adequate care.