Development and validation of classifiers and variable subsets for predicting nursing home admission

Predictors of nursing home admission in the older population in Belgium: a longitudinal follow-up of health interview survey participants

Background

This study examines predictors of nursing home admission (NHA) in Belgium in order to contribute to a better planning of the future demand for nursing home (NH) services and health care resources.

Methods

Data derived from the Belgian 2013 health interview survey were linked at individual level with health insurance data (2012 tot 2018). Only community dwelling participants, aged ≥65 years at the time of the survey were included in this study (n = 1930). Participants were followed until NHA, death or end of study period, i.e., December 31, 2018. The risk of NHA was calculated using a competing risk analysis.

Results

Over the follow-up period (median 5.29 years), 226 individuals were admitted to a NH and 268 died without admission to a NH. The overall cumulative risk of NHA was 1.4, 5.7 and 13.1% at respectively 1 year, 3 years and end of follow-up period. After multivariable adjustment, higher age, low educational attainment, living alone and use of home care services were significantly associated with a higher risk of NHA. A number of need factors (e.g., history of falls, suffering from urinary incontinence, depression or Alzheimer’s disease) were also significantly associated with a higher risk of NHA. On the contrary, being female, having multimorbidity and increased contacts with health care providers were significantly associated with a decreased risk of NHA. Perceived health and limitations were both significant determinants of NHA, but perceived health was an effect modifier on limitations and vice versa.

Conclusions

Our findings pinpoint important predictors of NHA in older adults, and offer possibilities of prevention to avoid or delay NHA for this population. Practical implications include prevention of falls, management of urinary incontinence at home and appropriate and timely management of limitations, depression and Alzheimer’s disease. Focus should also be on people living alone to provide more timely contacts with health care providers. Further investigation of predictors of NHA should include contextual factors such as the availability of nursing-home beds, hospital beds, physicians and waiting lists for NHA.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-022-03496-4.

Using machine learning for the personalised prediction of revision endoscopic sinus surgery

BACKGROUND

Revision endoscopic sinus surgery (ESS) is often considered for chronic rhinosinusitis (CRS) if maximal conservative treatment and baseline ESS prove insufficient. Emerging research outlines the risk factors of revision ESS. However, accurately predicting revision ESS at the individual level remains uncertain. This study aims to examine the prediction accuracy of revision ESS and to identify the effects of risk factors at the individual level.

METHODS

We collected demographic and clinical variables from the electronic health records of 767 surgical CRS patients ≥16 years of age. Revision ESS was performed on 111 (14.5%) patients. The prediction accuracy of revision ESS was examined by training and validating different machine learning models, while the effects of variables were analysed using the Shapley values and partial dependence plots.

RESULTS

The logistic regression, gradient boosting and random forest classifiers performed similarly in predicting revision ESS. Area under the receiving operating characteristic curve (AUROC) values were 0.744, 0.741 and 0.730, respectively, using data collected from the baseline visit until six months after baseline ESS. The length of time during which data were collected improved the prediction performance. For data collection times of 0, 3, 6 and 12 months after baseline ESS, AUROC values for the logistic regression were 0.682, 0.715, 0.744 and 0.784, respectively. The number of visits before or after baseline ESS, the number of days from the baseline visit to the baseline ESS, patient age, CRS with nasal polyps (CRSwNP), asthma, non-steroidal anti-inflammatory drug exacerbated respiratory disease and immunodeficiency or suspicion of it all associated with revision ESS. Patient age and number of visits before baseline ESS carried non-linear effects for predictions.

CONCLUSIONS

Intelligent data analysis identified important predictors of revision ESS at the individual level, such as the frequency of clinical visits, patient age, Type 2 high diseases and immunodeficiency or a suspicion of it.

Prevention Admission into Nursing homes (PAN): study protocol for an explorative, prospective longitudinal pilot study

BACKGROUND

In Switzerland, there is a lack of adequate rehabilitation services, and effective coordination, that take into account the multifactorial health risks of older people. The literature shows that the hospitalisation rate in rehabilitation facilities has increased in recent years and that a gender bias exists. Additionally, there is little or no evidence available on the effect that a post-acute care programme might have over an extended period on functioning, quality of life and the informal network of older people. Therefore, the aim of this trial is to evaluate the sustainability of post-acute care within three nursing homes in Zurich, Canton of Zurich, Switzerland.

METHODS

The Prevention Admission into Nursing homes (PAN) study is a explorative, prospective, longitudinal pilot trial based on a convenience sample of three long-term care facilities in the Swiss Canton of Zurich. The proposed pilot study will examine the effects of a post-acute care programme on people aged ≥65 years with a post-acute care potential ≥ three admitted to any of the three post-acute care units (n = 260). Older people of all sexes admitted to one of the post-acute care units and likely to be discharged to home within 8 weeks will be eligible for participation in the study. The primary endpoint is functionality based on the Barthel Index. The secondary endpoints are independency based on delirium, cognition, mobility, falling concerns, frailty, weight/height/body mass index, post-acute care capability, quality of life, and lastly, the informal network. As part of process evaluation, a qualitative evaluation will be conducted based on constructive grounded theory to specifically analyse how the experience of informal caregivers (n = 30) can contribute to a successful daily life 6 months after discharge.

DISCUSSION

We expect to observe improved functional status and independence after the post-acute care programme. The qualitative evaluation conducted with caregivers will complement our description of the transition of older people towards living at home.

TRIAL REGISTRATION

This study is registered in the German Clinical Trials Register under DRKS00016647 (registered on 23.05.2019).

Application Scenarios for Artificial Intelligence in Nursing Care: Rapid Review

Background

Artificial intelligence (AI) holds the promise of supporting nurses’ clinical decision-making in complex care situations or conducting tasks that are remote from direct patient interaction, such as documentation processes. There has been an increase in the research and development of AI applications for nursing care, but there is a persistent lack of an extensive overview covering the evidence base for promising application scenarios.

Objective

This study synthesizes literature on application scenarios for AI in nursing care settings as well as highlights adjacent aspects in the ethical, legal, and social discourse surrounding the application of AI in nursing care.

Methods

Following a rapid review design, PubMed, CINAHL, Association for Computing Machinery Digital Library, Institute of Electrical and Electronics Engineers Xplore, Digital Bibliography & Library Project, and Association for Information Systems Library, as well as the libraries of leading AI conferences, were searched in June 2020. Publications of original quantitative and qualitative research, systematic reviews, discussion papers, and essays on the ethical, legal, and social implications published in English were included. Eligible studies were analyzed on the basis of predetermined selection criteria.

Results

The titles and abstracts of 7016 publications and 704 full texts were screened, and 292 publications were included. Hospitals were the most prominent study setting, followed by independent living at home; fewer application scenarios were identified for nursing homes or home care. Most studies used machine learning algorithms, whereas expert or hybrid systems were entailed in less than every 10th publication. The application context of focusing on image and signal processing with tracking, monitoring, or the classification of activity and health followed by care coordination and communication, as well as fall detection, was the main purpose of AI applications. Few studies have reported the effects of AI applications on clinical or organizational outcomes, lacking particularly in data gathered outside laboratory conditions. In addition to technological requirements, the reporting and inclusion of certain requirements capture more overarching topics, such as data privacy, safety, and technology acceptance. Ethical, legal, and social implications reflect the discourse on technology use in health care but have mostly not been discussed in meaningful and potentially encompassing detail.

Conclusions

The results highlight the potential for the application of AI systems in different nursing care settings. Considering the lack of findings on the effectiveness and application of AI systems in real-world scenarios, future research should reflect on a more nursing care–specific perspective toward objectives, outcomes, and benefits. We identify that, crucially, an advancement in technological-societal discourse that surrounds the ethical and legal implications of AI applications in nursing care is a necessary next step. Further, we outline the need for greater participation among all of the stakeholders involved.

Chronic conditions and multimorbidity associated with institutionalization among Finnish community-dwelling older people: an 18-year population-based follow-up study

Aim

The aim of the study is to assess the association of chronic conditions and multimorbidity with institutionalization in older people.

Findings

Having dementia, mood or neurological disorder and/or five or more chronic conditions were associated with a higher risk of institutionalization.

Message

These risk factors should be recognized in primary care when providing and targeting care and support for home-dwelling older people.

Supplementary Information

The online version contains supplementary material available at 10.1007/s41999-021-00535-y.

Purpose

The ageing population is increasingly multimorbid. This challenges health care and elderly services as multimorbidity is associated with institutionalization. Especially dementia increases with age and is the main risk factor for institutionalization. The aim of this study was to assess the association of chronic conditions and multimorbidity with institutionalization in home-dwelling older people, with and without dementia.

Methods

In this prospective study with 18-year follow-up, the data on participants’ chronic conditions were gathered at the baseline examination, and of conditions acquired during the follow-up period from the municipality’s electronic patient record system and national registers. Only participants institutionalized or deceased by the end of the follow-up period were included in this study. Different cut-off-points for multimorbidity were analyzed. Cox regression model was used in the analyses. Death was used as a competing factor.

Results

The mean age of the participants (n = 820) was 74.7 years (64.0‒97.0). During the follow-up, 328 (40%) were institutionalized. Dementia, mood disorders, neurological disorders, and multimorbidity defined as five or more chronic conditions were associated with a higher risk of institutionalization in all the participants. In people without dementia, mood disorders and neurological disorders increased the risk of institutionalization.

Conclusion

Having dementia, mood or neurological disorder and/or five or more chronic conditions were associated with a higher risk of institutionalization. These risk factors should be recognized when providing and targeting care and support for older people still living at home.

Supplementary Information

The online version contains supplementary material available at 10.1007/s41999-021-00535-y.

Subjective and objective health predicting mortality and institutionalization: an 18-year population-based follow-up study among community-dwelling Finnish older adults

BACKGROUND

Objective health measures, such as registered illnesses or frailty, predict mortality and institutionalization in older adults. Also, self-reported assessment of health by simple self-rated health (SRH) has been shown to predict mortality and institutionalization. The aim of this study was to assess the association of objective and subjective health with mortality and institutionalization in Finnish community-dwelling older adults.

METHODS

In this prospective study with 10- and 18-year follow-ups, objective health was measured by registered illnesses and subjective health was evaluated by simple SRH, self-reported walking ability (400 m) and self-reported satisfaction in life. The participants were categorized into four groups according to their objective and subjective health: 1. subjectively and objectively healthy, 2. subjectively healthy and objectively unhealthy, 3. subjectively unhealthy and objectively healthy and 4. subjectively and objectively unhealthy. Cox regression model was used in the analyses. Death was used as a competing factor in the institutionalization analyses.

RESULTS

The mean age of the participants (n = 1259) was 73.5 years (range 64.0-100.0). During the 10- and 18-year follow-ups, 466 (37%) and 877 (70%) died, respectively. In the institutionalization analyses (n = 1106), 162 (15%) and 328 (30%) participants were institutionalized during the 10- and 18-year follow-ups, respectively. In both follow-ups, being subjectively and objectively unhealthy, compared to being subjectively and objectively healthy, was significantly associated with a higher risk of institutionalization in unadjusted models and with death both in unadjusted and adjusted models.

CONCLUSIONS

The categorization of objective and subjective health into four health groups was good at predicting the risk of death during 10- and 18-year follow-ups, and seemed to also predict the risk of institutionalization in the unadjusted models during both follow-ups. Poor subjective health had an additive effect on poor objective health in predicting mortality and could therefore be used as part of an older individual's health evaluation when screening for future adverse outcomes.

A longitudinal study of the impact of social network size and loneliness on cognitive performance in depressed older adults

Objectives: To examine the association of social network size and loneliness with cognitive performance and -decline in depressed older adults.Method: A sample of 378 older adults [70.7 (7.4) years] with a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition diagnosis of current depressive disorder were recruited from primary care and specialized mental health care. Cognitive performance was assessed at baseline and 2 years follow-up with the Stroop colored-word test, a modified version of the Auditory Verbal Learning Task and the Digit Span subtest from the Wechsler Adult Intelligence Scale, encompassing four cognitive domains; processing speed, interference control, memory, and working memory. Social network size was assessed with the Close Person Inventory and loneliness with the de Jong Gierveld Loneliness Scale at baseline.Results: After adjusting for baseline working memory performance, loneliness was associated with impaired working memory after 2 years [B = -0.08 (-0.17 to 0.00)]. This association was no longer significant after adjusting for age, sex, education level, physical activity, alcohol use and depressive symptom severity [B = -0.07 (-0.16 to 0.03)]. A backward elimination procedure revealed education level to be the only covariable to explain this association. Loneliness was not associated with impairments or decline in other cognitive domains. Social network size was not associated with cognitive impairments or decline.Conclusion: Social network size and loneliness do not predict cognitive decline in depressed older adults.

Health and care service utilisation and cost over the life-span: a descriptive analysis of population data

BACKGROUND

Current demographic changes affect both the level and composition of health and care needs in the population. The aim of this study was to estimate utilisation and cost for a comprehensive range of health and care services by age and gender to provide an in-depth picture of the life-span pattern of service needs and related costs.

METHODS

Data on service use in 2010 for the entire population in Norway were collected from four high-quality national registers. Cost for different services were calculated combining data on service utilisation from the registries and estimates of unit cost. Data on cost and users were aggregated within four healthcare services and seven long-term care services subtypes. Per capita cost by age and gender was decomposed into user rates and cost per user for each of the eleven services.

RESULTS

Half of the population is under 40 years of age, but only a quarter of the health and care cost is used on this age group. The age-group of 65 or older, on the other hand, represent only 15% of the population, but is responsible for almost half of the total cost. Healthcare cost dominates in ages under 80 and mental health services dominates in adolescents and young adults. Use of other healthcare services are high in middle aged and elderly but decreases for the oldest old. Use of care services and in particular institutional care increases in old age. Healthcare cost per user follows roughly the same age pattern as user rates, whereas user cost for care services typically are either relatively stable or decrease with age among adults. Gender differences in the age pattern of health and care costs are also revealed and discussed.

CONCLUSION

The type of services used, and the related cost, show a clear life-span as well as gender pattern. Hence, population aging and narrowing gender-gap in longivety calls for high policy awarness on changing health and care needs. Our study also underscores the need for an attentive and pro-active stance towards the high service prevalence and high cost of mental health care in our upcoming generations.

Developing and validating models for predicting nursing home admission using only RAI-HC instrument data

OBJECTIVE

In recent years research has identified important predictors for nursing home admission (NHA). However, as far as we know, the previous risk models use complex variable sets from many sources and the output is a single risk value. The objective of this study was to develop an NHA risk model with a variable set from single data source and richer output information.

METHODS

In this study, we developed a model selecting variables only from the RAI-HC (Resident Assessment Instrument - Home Care) system. Furthermore, we used principal component analysis and K-means clustering to target proper interventions for high-risk clients.

RESULTS

The performance of the model was close to the complex previous model (recall [Formula: see text] vs. [Formula: see text] and specificity [Formula: see text] vs. [Formula: see text]). For the risk clients, three intervention clusters (deficiency in physical functionality, deficiency in cognitive functionality and depression and mood disorders) were found.

CONCLUSION

The NHA risk model and intervention clusters are important because they enable the identification of proper interventions for the right clients. The fact that the model with RAI-HC data alone was accurate enough simplifies the integration of the NHA risk model into practice because it uses data from one system and the algorithm can be integrated easily into the source system.

A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models

OBJECTIVES

The objective of this study was to compare performance of logistic regression (LR) with machine learning (ML) for clinical prediction modeling in the literature.

STUDY DESIGN AND SETTING

We conducted a Medline literature search (1/2016 to 8/2017) and extracted comparisons between LR and ML models for binary outcomes.

RESULTS

We included 71 of 927 studies. The median sample size was 1,250 (range 72-3,994,872), with 19 predictors considered (range 5-563) and eight events per predictor (range 0.3-6,697). The most common ML methods were classification trees, random forests, artificial neural networks, and support vector machines. In 48 (68%) studies, we observed potential bias in the validation procedures. Sixty-four (90%) studies used the area under the receiver operating characteristic curve (AUC) to assess discrimination. Calibration was not addressed in 56 (79%) studies. We identified 282 comparisons between an LR and ML model (AUC range, 0.52-0.99). For 145 comparisons at low risk of bias, the difference in logit(AUC) between LR and ML was 0.00 (95% confidence interval, -0.18 to 0.18). For 137 comparisons at high risk of bias, logit(AUC) was 0.34 (0.20-0.47) higher for ML.

CONCLUSION

We found no evidence of superior performance of ML over LR. Improvements in methodology and reporting are needed for studies that compare modeling algorithms.

Prevalence and persistent use of psychotropic drugs in older adults receiving domiciliary care at baseline

Background

Little is known about the use of psychotropic drugs in older adults receiving domiciliary care. The first aim was to describe the prevalence and persistency of use of psychotropic drugs in older adults (≥ 70 years) with and without dementia receiving domiciliary care. Furthermore, the second aim was to explore factors associated with persistent drug use at two consecutive time-points. Lastly, we aimed to examine if use of psychotropic drugs changed after admission to a nursing home.

Methods

In total, 1001 community-dwelling older adults receiving domiciliary care at inclusion participated in the study. Information about psychotropic drug use was collected at baseline, after 18 months and after 36 months. The participants’ cognitive function, neuropsychiatric symptoms (NPS) and physical health were assessed at the same assessments. Participants were evaluated for dementia based on all gathered information. Formal level of care (domiciliary care or in a nursing home) was registered at the follow-up assessments.

Results

Prevalence and persistent use of psychotropic drugs in older adults receiving domiciliary care was high. Participants with dementia more often used antipsychotics and antidepressants than participants without dementia. The majority of the participants using antipsychotic drugs used traditional antipsychotics. Younger age was associated with higher odds for persistent use of antipsychotics and antidepressants, and lower odds for persistent use of sedatives. Severity of NPS was associated with persistent use of antidepressants. The odds for use of antipsychotics and antidepressants were higher in those admitted to a nursing home as compared to the community-dwelling participants at the last follow-up.

Conclusion

There was a high prevalence and persistency of use of psychotropic drugs. The prevalence of use of traditional antipsychotics was surprisingly high, which is alarming. Monitoring the effect and adverse effects of psychotropic drugs is an important part of the treatment, and discontinuation should be considered when possible due to the odds for severe adverse effects of such drugs in people with dementia.

Electronic supplementary material

The online version of this article (10.1186/s12877-019-1126-y) contains supplementary material, which is available to authorized users.

What factors influence older people's intention to enrol in nursing homes? A cross-sectional observational study in Shanghai, China

OBJECTIVES

Given the increasing need of long-term care and the low occupancy rate of nursing homes in Shanghai, this study attempts to explore what factors influence older people's intention to enrol in nursing homes.

DESIGN

A cross-sectional observational study based on the theory of reasoned action was conducted. Survey data were collected from subjects during face-to-face interviews. Structural equation modelling was employed for data analysis.

SETTING

This study was conducted in six community health service centres in Shanghai, China. Two service centres were selected in urban, suburban and rural areas, respectively.

PARTICIPANTS

A total of 641 Shanghai residents aged over 60 were surveyed.

RESULTS

Structural equation modelling analysis showed that the research model fits the data well (χ²/df=2.948, Comparative Fit Index=0.972 and root mean squared error of approximation =0.055). Attitude (β=0.41, p<0.01), subjective norm (β=0.28, p<0.01) and value-added service (β=0.16, p<0.01) were directly associated with enrolment intention, explaining 32% of variance in intention. Attitude was significantly influenced by loneliness (β=-0.08, p<0.05), self-efficacy (β=0.32, p<0.01) and stigma (β=-0.24, p<0.01), while subjective norm was significantly influenced by life satisfaction (β=-0.15, p<0.01) and stigma (β=-0.43, p<0.01).

CONCLUSIONS

This study advances knowledge regarding the influencing factors of older people's intention to enrol in nursing homes. It suggests that Chinese older persons' perceived stigma has the strongest indirect effect on their intention to enrol in nursing homes. This is unique to the Chinese context and has practical implications for eldercare in China and other Asian countries with similar sociocultural contexts.