Hospitalized older adult: predictors of functional decline

Objective: to identify the predictors of functional decline in hospitalized individuals aged 70 or over, between: baseline and discharge; discharge and follow-up, and baseline and three-month follow-up. Method: a prospective cohort study conducted in internal medicine services. A questionnaire was applied (clinical and demographic variables, and predictors of functional decline) at three moments. The predictors were determined using the binary logistic regression model. Results: the sample included 101 patients, 53.3% female, mean age of 82.47 ± 6.57 years old. The predictors that most contributed to decline in hospitalization were the following: previous hospitalization (OR=1.8), access to social support (OR=4.86), cognitive deficit (OR=6.35), mechanical restraint (OR=7.82), and not having a partner (OR=4.34). Age (OR=1.18) and medical diagnosis (OR=0.10) were the predictors between discharge and follow-up. Being older, delirium during hospitalization (OR=5.92), and presenting risk of functional decline (OR=5.53) were predictors of decline between the baseline and follow-up. Conclusion: the most relevant predictors were age, previous hospitalization, cognitive deficit, restraint, social support, not having a partner, and delirium. Carrying out interventions aimed at minimizing the impact of these predictors can be an important contribution in the prevention of functional decline.


Method
In order to identify the largest number of predictors of FD, a questionnaire was carried out which included sociodemographic and clinical data, as well as different scales (Table 1).
The assessment of the functional capacity was performed with the Katz Index (KI). It is an instrument with six basic activities of daily life (BADL) of dichotomous reply (0=dependent, 1=independent). The older adults and/or caregivers were asked to describe their functional capacity prior to hospitalization (baseline), reporting the last two weeks. The decline in the BADL between baseline and discharge was defined as t0; t1 corresponded to the decline between discharge and follow-up (FU); and t2 reported the decline between baseline and FU ( Figure 1). In this study, FD was defined as any decline in one or more points in the KI between the three moments at which the assessment was conducted. In the initial assessment, the patients admitted in the services were analyzed daily in order to identify the ones that presented eligibility criteria. The data were collected by the researchers through: heterofilling of the questionnaire (preferably with the patient or, if not possible, with the informal caregivers, or with the health team: nurses, physicians, and operational assistants), consultation of the clinical diary, of the electronic medical record, and through telephone contact to obtain the followup data. A three-month FU was chosen given that some studies reported that, in this period (1 st and 3 rd months after discharge), a significant number of older adults can recover their functionality (30)(31) .
In the analysis of the data, techniques of descriptive and inferential statistics were used. In the univariate analysis, the Student's t test was used (when the normality of the distribution was not verified, Mann-Whitney's U test was used), as well as ANOVA (when the normality of the distribution was not verified, Kruskal-Wallis H test was used), and the chi-square test with Odds Ratio (OR) (when the estimates of the chi-square test were not verified, Fisher's exact test was used).
In the multivariate analysis, the binary logistic regression model was used for the three declines under study (t0, t1, and t2). The variables to include in the model were selected in accordance with the following: the number of cases (one predictor for every ten cases), p≤0.15, and predictors with evidence in the literature (32)(33)(34) .
The correlation between the independent variables was

Results
A total of 117 patients were included, of which the following were excluded: 10 due to death, four due to transfers to other services, one due to absence of baseline, and one for hospitalization of less than   (26) Admission and during hospitalization Restraint Observation grid of physical restriction (27) During hospitalization Risk of functional decline ISAR-HP -Portuguese version (28) Admission Functional capacity of the OA Katz Index (29) Admission, discharge and follow-up    compared to the patients who did not decline), those who were more frequently hospitalized in the last year, those who had CCI and ERRD with higher scores, and those who had longer hospitalizations (two more days -median).
They were also those who reported feeling sad or depressed more frequently and had delirium episodes during hospitalization. It was observed that functional decline at t2 is dependent on the risk assessment by ISAR-HP, and that the older adults at risk are 10.72 times more likely to have functional decline.
The use of restraints is worsened during hospitalization and is linked to other predictors (41) , such as cognitive deficit, a predisposing factor for delirium which, in turn, can predetermine the use of restraints. This hospital practice is more frequent in older adults, who are already vulnerable to the adverse events resulting from restraints (e.g., incontinence, isolation, pressure injury, infection, delirium, and death due to aspiration) (41-42) .
The use of alternative measures to the restraints must be can be independent of it (30) , with hospitalization playing a very significant role (15) . The fifth limitation is that the assessment during hospitalization took place at a single moment. Extending this assessment to more moments could allow for the identification of a greater number of predictors. Finally, these results must be generalized with caution, taking into account the specific context (internal medicine services) in a university hospital. Multicenter studies would be decisive to determine the predictors of functional decline among hospitalized older adults in Portugal.
This study allows evidencing the factors that most contributed to functional decline in the older adults hospitalized in internal medicine services, with the following standing out: age, hospitalization in the last year, cognitive deficit, restraint, and delirium. Some of these predictors stem from the health-disease process; however, others can be associated with the hospital practice.

Conclusion
This study showed that the transversal predictors to FD were as follows: advanced age, hospitalization in the last year, delirium, and having an assessment of risk of decline. During hospitalization, the most significant predictors of FD were the following: previous hospitalization, access to social support, cognitive deficit, mechanical restraint, and not having a partner. Between discharge and follow-up, the main predictors of FD were age and medical diagnosis. Finally, between baseline and follow-up, advanced age, delirium, and the risk of FD were the most relevant predictors. Implementing targeted interventions with specific guidelines for these predictors, especially those associated with the hospital practice, such as the use of restraints and delirium, will certainly contribute to preventing FD in hospitalized older adults.