Nursing Risk Assessment for the Elderly and Its Data-Driven Optimization Model

China is one of the fastest growing countries in the world in terms of aging process, the proportion of elderly people over 65 years of age in the total population has increased from 7% to 14%, while developed countries took decades or even hundreds of years, China only took 27 years and will maintain a high incremental rate for a long time to come [1-3]. The accelerated rate of aging has also led to an increase in chronic non-communicable diseases (NCDs) in the elderly. Caring for the elderly with chronic NCDs is a long-term process, and with the progression of the disease, the difficulty of care increases, the task is aggravated, and the caregiver’s physiological and psychological health and economic status are affected [4-6]. Elderly groups, as the nation’s builders and pioneers, need to be provided with better care during the process of medical treatment. As patients are old and their physical functions tend to deteriorate, they are prone to various risk events in nursing care, which may jeopardize their health and delay the process of rehabilitation and discharge [7-9]. Therefore, it is necessary to carry out nursing risk assessment, scientific assessment around the patient’s own reality, early warning of potential risks, and accordingly provide targeted care to reduce the rate of adverse events [10-12].

Nursing risk refers to all the unsafe events that may occur in the clinical nursing work, from the identification of risk, assessment, treatment, evaluation and management, can effectively avoid the risk. Nursing risk management started late in China and lacks risk management and a unified evaluation system from the management level [13-15]. Prospective management is a management model that foresees quality risks and takes prior control [16-17]. Much of modern nursing management adopts the concept of prospective management, moving the risk management link forward and proposing risk prevention measures, so as to ensure the safe operation of nursing. The risk assessment without the support of information system has the defects of too much assessment content, too complicated, insufficient intelligence, and large workload [18-20].

In this study, an optimization model that can be used for nursing risk assessment in elderly population was designed. Firstly, the selection of research subjects was carried out according to the relevant criteria to determine the appropriate research sample. The study referred to questionnaires such as Big Five Personality and Suggested Response, and CGRA Geriatric Nursing Risk Scale was selected as the nursing risk assessment tool. After completing data processing and entry, SPSS 18.0 statistical software was used to analyze the data, and the risk assessment model used was a dichotomous logistic regression model. The effects of gender and age on nursing risk were tested separately during the assessment process. Subsequently, a one-way ANOVA was conducted to test the significance of the effect of CGRA score, visual acuity, and chronic diseases on nursing risk, and based on this, a multifactorial analysis was carried out to determine the risk factors and protective factors of nursing risk. Finally, a graded care method based on nursing risk assessment was proposed to grade care for elderly groups with different nursing risks, and experiments were designed to test the effect of graded care on nursing risk reduction.

2

Data-driven research design for assessing caregiving risk in the elderly population

2.1

Purpose of the study

Through quantitative survey to understand and analyze the current situation and influencing factors of the risk of caregiving for the elderly group in long-term care facilities, and initially construct a risk assessment model of caregiving for the elderly group caregivers, to provide a reference basis for the early identification and early intervention of the risk of caregiving for the elderly group caregivers.

2.2

Research design

2.2.1

Research Objectives

A total of 305 caregivers from 8 long-term care facilities in X city were selected by convenience sampling from July to December 2023 for this study. 1)

Inclusion criteria for study subjects: (1)

Caregivers were aged ≥55 years.

(2)

The caregivers were professionals who had been educated and trained to provide paid caregiving services in relation to the disease.

(3)

The caregiver had continuously cared for the older adult for ≥1 month.

(4)

The caregiver is ≥18 years old.

(5)

The caregiver is conscious and has normal language and communication skills.

(6)

The caregiver gives informed consent and participates in this survey voluntarily.

2)

Exclusion criteria of research subjects: (1)

Non-employed caregivers such as family members, friends and social workers.

(2)

Non-professionals such as babysitters who have not received education and training on relevant disease knowledge.

3)

The sample size method is as follows:

Sample size $= (t_{α}^{2} P Q / d^{2}) * [1 + (10 % ~ 20 %)]$ $$ = \left( {{t_\alpha }^2PQ/{d^2}} \right)^*\left[ {1 + (10\% \sim 20\% )} \right]$$, where α = 0.05 (two-sided test), t_α = 1.96, P are the incidence of nursing risk, Q = 1 − P, d are the allowable error. The value of P is taken as 0.532, and the value of d is defined as 0.05. The values are brought into the formula and 10% to 20% of the sample is considered, resulting in a minimum sample size of 225 to 254 cases.

2.2.2

Research tools

1)

Big Five Personality Questionnaire

The questionnaire was developed according to the Big Five personality OCEAN theory, including 5 dimensions of neuroticism, extraversion, openness, rigor and agreeableness, with a total of 60 items, each dimension was composed of 12 personality factors, and the average reliability of the 5 subscales was 0.75. Subjects responded to the scale, selecting “strongly disagree”, “disagree”, “neutral”, “agree” and “strongly agree”, and scored “1” to “5” in turn. The score of the 5-dimensional scale is 12~60 points, of which 60 points represent the highest score of the personality factor, and 12 points represent the lowest score of the personality factor. The scale has been applied to the elderly population and has been shown to have good reliability and construct validity [21].

2)

Simple Response Questionnaire

This questionnaire is used to assess an individual’s coping style. The questionnaire has a total of 20 items, including two dimensions: positive coping and negative coping. Positive coping reflects the characteristics of individual positive coping and consists of items 1~12. Negative coping reflects the characteristics of individual negative coping, which is composed of items 13~20, and the Cronbach’s α coefficient of the questionnaire is 0.90. The questionnaire selects “not used” to record “0 points”, “occasionally used” to record “1 point”, sometimes “to use” 2 points, and “often used” to record “3 points”, and the results were the average score of the positive coping dimension and the average score of the negative coping dimension, the higher the positive score, the higher the respondent tended to respond positively, and the higher the negative score, the higher the negative score, the more inclined to negative coping.

3)

CGRA Geriatric Care Risk Scale

This questionnaire is used to assess the care risk profile of the elderly in the past 1 month. The questionnaire consists of 15 items, half positive and half negative. The most commonly used scoring method is the “0-0-1-1” scoring method, that is, the items that choose “no risk at all” or “low risk” are all recorded as “0 points”, and the items that choose “medium risk” or “high risk” are all recorded as “1 point”, so the total score range of 15 items is 0~15 points, and the higher the score of the scale, the higher the nursing risk. Its Cronbach’s α coefficient was applied to the elderly group of 0.75, respectively.

2.2.3

Research methodology

The researcher himself, after obtaining the consent of the administrator of the organization and obtaining the informed consent of the respondents, conducted the survey by distributing the paper version of the questionnaire in a face-to-face manner. Unified guidelines were used during the survey, and the investigator explained or read the questionnaire on the spot for those who had doubts or reading difficulties, and for those who had difficulties in filling out the questionnaire, the investigator could assist in filling out the questionnaire. The questionnaires were collected on the spot and checked for missing items.

2.2.4

Statistical methods

Data were checked by double entry and analyzed using SPSS 18.0 statistical package. 1)

The general information of caregivers and geriatric groups was described by frequency, percentage, and mean ± standard deviation.

2)

Geriatric group caregiving risk scores were described by mean ± standard deviation, and correlation analyses for each scale were performed using Spearman’s correlation.

3)

Multifactorial analyses of potential influences on nursing risk in the geriatric population were analyzed by logistic regression.

4)

Logistic stepwise regression analysis was used for model construction. Significance levels of α = 0.05, P < 0.05 represent statistically significant differences.

2.2.5

Quality control

1)

The survey was optimized in terms of the simplicity and readability of the questionnaire through a pre-survey, and caregivers who met the inclusion and exclusion criteria were explicitly used as survey respondents before implementation.

2)

Prior to the survey, respondents were given a detailed understanding of the questionnaire content, survey techniques, and precautions to ensure the completeness and accuracy of quantitative data collection. The purpose of this data collection and the principle of confidentiality were emphasized to the survey respondents during the survey to eliminate their doubts and ensure the reliability of the collected data. After the survey, the investigator promptly recovered the questionnaire and reviewed it to ensure the validity of the questionnaire results. The questionnaire screening criteria for answer similarity and omission of the topic more than 15%.

3)

Due to the sensitivity of the subject matter of this study, in order to ensure that the data collection is true and effective, the researcher emphasizes the purpose of this study and the principle of confidentiality during the research process, while the researcher establishes a good relationship with the caregivers and strives to ensure that the caregivers fill out the questionnaire truthfully.

2.2.6

Ethical principles

This study was conducted with the consent of the administrators of the long-term care facility and the informed consent of the caregivers of the elderly group. The purpose and significance of the study and the principle of confidentiality were emphasized during the conduct of the survey.

3

Logistic regression-based optimization model for nursing risk assessment

3.1

Logistic regression model and its application

3.1.1

Logistic regression models

Logistic regression is a kind of generalized linear regression, which is suitable for categorical prediction where the dependent variable is dichotomous, or the probability of something occurring, and is widely used in medicine. Logistic regression can be used in a number of areas such as analysis of risk factors for some diseases, dose-response studies of medications, and analysis of prognosis of diseases, etc. There are various types of logistic regression models, which can be classified into dichotomous regression models and multicategorical regression models depending on the Logistic regression models are of various types, which can be divided into dichotomous regression models and multichotomous regression models according to the dependent variable, and the variables can be continuous or discrete. In this paper, we mainly use dichotomous logistic regression model. The definition of regression model is as follows:

Let P be the incidence of nursing risk, this dependent variable has only 2 outcomes (occurrence and non-occurrence) and it has m independent variables x₁, x₂, ⋯, x_m, then the expression of Logistic regression can be divided into: 1 $P = \frac{\exp (β_{0} + β_{1} x_{1} + β_{2} x_{2} + \dots + β_{m} x_{m})}{1 + \exp (β_{0} + β_{1} x_{1} + β_{2} x_{2} + \dots + β_{m} x_{m})}$ $$P = \frac{{\exp \left( {{\beta_0} + {\beta_1}{x_1} + {\beta_2}{x_2} + \ldots + {\beta_m}{x_m}} \right)}}{{1 + \exp \left( {{\beta_0} + {\beta_1}{x_1} + {\beta_2}{x_2} + \ldots + {\beta_m}{x_m}} \right)}}$$ 2 $\frac{P}{1 - P} = \exp (β_{0} + β_{1} x_{1} + β_{2} x_{2} + \dots + β_{m} x_{m})$ $$\frac{P}{{1 - P}} = \exp \left( {{\beta_0} + {\beta_1}{x_1} + {\beta_2}{x_2} + \ldots + {\beta_m}{x_m}} \right)$$

Taking logarithms on both sides, the general form of the bicategorical logistic can be expressed as: 3 $\log i t (P) = \ln (\frac{P}{1 - P}) = β_{0} + β_{1} x_{1} + β_{2} x_{2} + \dots + β_{m} x_{m}$ $$\log it(P) = \ln \left( {\frac{P}{{1 - P}}} \right) = {\beta_0} + {\beta_1}{x_1} + {\beta_2}{x_2} + \cdots + {\beta_m}{x_m}$$

The model is similar to a linear regression model, but it does a transformation of the incidence on the left side, i.e., it transforms to $\ln (\frac{P}{1 - P})$ $$\ln \left( {\frac{P}{{1 - P}}} \right)$$, rather than the dependent variable itself. It captures the effect of the m independent variables on the outcome, and the magnitude of the effect of each independent variable can be expressed by the corresponding regression coefficient β₀, β₁, β₂, ⋯, β_m in the model [22].

The coefficient β of the independent variable in the logistic regression model is related to the dominance ratio (OR) in epidemiology as: 4 $O R = e^{β}$ $$OR = {{\text{e}}^\beta }$$

OR is the OR value, which means that the OR value is closely related to the value of β. The larger the β, the larger the OR value.

OR values are often used in comparative studies of the risk of a number of diseases, and various types of interpretations are often made in logistic regression based on OR values rather than regression coefficients. In practice, OR values are often interpreted as the risk incidence of these factors.

3.1.2

Logistic Regression Modeling in Health Care

Logistic regression is often used in some data analysis or predictive modeling. Logistic regression is more widely used in healthcare data prediction, such as heart disease, high blood pressure prediction, etc. Logistic regression is very widely used in healthcare applications, it is a relatively reliable model, and because of the special function of its regression with OR value, it can give healthcare personnel the analysis and research on the attributes of risk susceptibility while predicting the risk of nursing care [23]. Therefore, the use of logistic regression for nursing risk prediction is more comprehensive. In this paper, Logistic regression is used as the base regression model for this model.

3.2

Logistics Geriatric Nursing Risk Assessment Model Application

3.2.1

Incidence of risks in the care of older persons in nursing facilities

Based on the definition of nursing risk in the scale used in the previous section, the current status of nursing risk occurrence in older adults in the past 1 year was investigated, and the incidence of nursing risk in older adults in nursing facilities is shown in Table 1. There were 141 cases of nursing risk in 305 older adults, and the incidence of nursing risk was 46.2%. Of these, 83 older adults had one nursing risk, with a one-time nursing risk rate of 27.2%, and 58 older adults had multiple nursing risks, with a nursing risk rate of 19%.

Table 1.

The incidence of nursing risk in elderly care institutions

The number of risk of nursing	Sample size (n=305)	Nursing risk frequency	Incidence rate
1	305	83	27.2%
≥2	305	58	19%
Total	305	141	46.2%

Table 2 shows the comparison of the incidence of nursing risk among the elderly in different gender and age care facilities. This study showed that out of 141 cases of elderly at nursing risk, 58 were male and 83 were female. The incidence of nursing risk was 41.1% and 58.9% in males and females respectively. Primary care risk was 17 cases (12.06%) in males and 30 cases (21.28%) in females. Multiple nursing risks were 41 (29.08%) in males and 53 (37.59%) in females. The incidence of nursing risks was higher in females than in males, and the difference was statistically significant (P < 0.05). The incidence of nursing risk increased with age (Trend chi-square test χ²= 22.687, P < 0.01). The survey showed that the incidence of 1 care risk and multiple care risk increased with age in the elderly.

Table 2.

Comparison of the incidence of elderly nursing risk with each gender and age

Item		Number of risks=1		χ²	P
Item		Count	Percentage	χ²	P
Gender	Male	17	12.06%	14.338	0.002
Gender	Female	30	21.28%	14.338	0.002
Age	[60,70)	4	2.84%	22.687	0.001
	[70,80)	15	10.64%
	[80,90)	62	43.97%
	≥90	3	2.13%
Item		Number of risks≥2		χ²	P
Item		Count	Percentage	χ²	P
Gender	Male	41	29.08%	14.054	0.002
Gender	Female	53	37.59%	14.054	0.002
Age	[60,70)	2	1.42%	23.387	0.000
	[70,80)	12	8.51%
	[80,90)	39	27.66%
	≥90	4	2.84%

3.2.2

Risk factor analysis of caregiving risks for older people in different situations

1)

Univariate analysis of nursing risk

Table 3 shows the analysis of factors affecting the risk of caregiving among the elderly in nursing facilities. The study showed that gender, age, CGRA score, duration of stay in the nursing facility, vision, chronic diseases, number of illnesses, number of long-term medications, antihypertensive medications, cardiac medications, anti-anxiety medications, static balance, dynamic balance, physical fitness status, regularity of life, physical activity, clutter accumulation in the aisles, furniture fixation, use of walkers by the elderly, and the presence of a fixed handrail in the bathrooms were found to be significant by the χ²-test (p < 0.05).

2)

Multifactorial analysis of nursing risk

The assignment of the regression analysis variables and the results of the analysis are shown in Tables 4 and 5, respectively. With the occurrence of nursing risk events as the dependent variable, all factors related to nursing risk were included in the logistic regression model for multifactorial analysis.

The multifactorial logistic regression analysis showed that gender, age, CGRA score, visual acuity, long-term medication, dynamic balance, whether the furniture was fixed or not, whether the bathroom had fixed handrails or not, and chronic diseases were risk factors for nursing care risk in the elderly. While physical activity had a negative regression coefficient and was a protective factor for elderly care risk.

Table 3.

Influencing factors of nursing risk occurrences in elderly people

Danger factors		n	Risk occurrences	Percentage	χ²	P
Gender	Male	62	40	64.5%	9.122	0.000
Gender	Female	79	28	35.4%	9.122	0.000
Age	[60,70)	14	4	28.6%	16.564	0.001
	[70,80)	38	14	36.8%
	[80,90)	79	44	55.7%
	≥90	10	5	50.0%
CGRA evaluation	[45,75)	25	8	32.0%	20.109	0.000
	[75,105)	61	30	49.2%
	≥105	55	28	50.9%
Check-in time	[1,5)	109	50	45.9%	7.930	0.025
	[5,11)	19	11	57.9%
	≥11	13	6	46.2%
Sight	Normal	66	29	43.9%	9.046	0.000
Sight	Abnormal	75	35	46.7%	9.046	0.000
Chronic disease	Yes	107	53	49.5%	16.513	0.001
Chronic disease	No	34	10	29.4%	16.513	0.001
Number of chronic diseases	≥3	41	23	56.1%	20.133	0.000
	≤3	77	35	45.5%
	Null	23	5	21.7%
Long-term dose	≥3	108	51	47.2%	18.247	0.002
	≤3	9	5	55.6%
	Null	24	7	29.2%
Anti-hypertensive	Yes	82	41	50.0%	10.760	0.003
Anti-hypertensive	No	59	23	39.0%	10.760	0.003
Heart drug	Yes	50	27	54.0%	10.744	0.000
Heart drug	No	91	39	42.9%	10.744	0.000
Antianxiety drug	Yes	4	3	75.0%	4.610	0.036
Antianxiety drug	No	137	65	47.4%	4.610	0.036
Static balance	Normal	95	46	48.4%	9.845	0.002
Static balance	Abnormal	46	20	43.5%	9.845	0.002
Physical ability	Normal	84	31	36.9%	16.147	0.000
Physical ability	Abnormal	57	31	54.4%	16.147	0.000
Dynamic balance	Normal	87	34	39.1%	21.362	0.001
Dynamic balance	Abnormal	54	33	61.1%	21.362	0.001
Regular routine	Yes	128	52	40.6%	35.189	0.000
Regular routine	No	13	11	84.6%	35.189	0.000
Physical exercise	Yes	82	34	41.5%	6.917	0.031
Physical exercise	No	59	32	54.2%	6.917	0.031
Cluttered storage in passage	Yes	99	45	45.5%	7.413	0.033
Cluttered storage in passage	No	42	22	52.4%	7.413	0.033
Furniture fixture	Yes	27	8	29.6%	14.663	0.000
Furniture fixture	No	114	59	51.8%	14.663	0.000
Using walking aid	Yes	84	41	48.8%	14.941	0.000
Using walking aid	No	57	21	36.8%	14.941	0.000
Grib bar installation	Yes	51	21	41.2%	5.669	0.027
Grib bar installation	No	90	47	5%	5.669	0.027

Table 4.

Logistic regression analysis of variable assignment

	Variable	Valuation
Y	Nursing risk	Occurs=1	Not ours=0
X₁	Gender	Male=0	Female=1
X₂	Age	60-69=1	70-79=2	80-89=3	≥90=4
X₃	CGRA evaluation	45-74=1	75-104=2	≥105=4
X₄	Check-in time	1-5 year=1	6-10 year=2	≥11 year=3
X₅	Sight	Normal=0	Abnormal=1
X₆	Chronic disease	No =0	Yes =1
X₇	Number of chronic diseases	≥3=0	<3=1	Null=2
X₈	Long-term dose	No =0	<3 =1	≥3=2
X₉	Anti-hypertensive	No =0	Yes =1
X₁₀	Heart drug	No =0	Yes =1
X₁₁	Antianxiety drug	No =0	Yes =1
X₁₂	Static balance	Yes =0	No =1
X₁₃	Physical ability	Yes =0	No =1
X₁₄	Dynamic balance	Yes =0	No =1
X₁₅	Regular routine	Yes =0	No =1
X₁₆	Physical ability	Yes =0	No =1
X₁₇	Physical exercise	No =0	Yes =1
X₁₈	Cluttered storage in passage	Yes =0	No =1
X₁₉	Furniture fixture	Yes =0	No =1
X₂₀	Using walking aid	No =0	Yes =1
X₂₁	Grib bar installation	Yes =0	No =1

Table 5.

Logistic regression analysis of multiple factors of nursing risk in the elderly

	B	S.E.	Wald	P	Exp(B)	C.I.(95%)
	B	S.E.	Wald	P	Exp(B)	Lower	Upper
Gender	0.750	0.281	7.698	0.005	2.180	1.358	3.361
Age	0.524	0.151	13.251	0.001	1.677	1.264	88
CGRA evaluation	0.538	0.159	14.635	0.002	1.708	1.260	98
Sight	0.463	0.228	4.726	0.004	1.613	1.051	2.385
Long-term dose	0.511	0.214	4.742	0.02	1.574	1.045	2.372
Dynamic balance	0.882	0.393	5.008	0.02	2.379	1.122	4.900
Furniture fixture	1.538	0.409	15.626	0.023	4.476	2.175	9.409
Using walking aid	0.671	0.221	10.092	0.000	1.978	1.204	2.919
Grib bar installation	0.668	0.239	5.378	0.018	1.869	1.109	3.323
Chronic disease	0.534	0.209	11.061	0.001	1.782	1.273	2.304
Physical exercise	-0.755	0.352	4.720	0.032	0.507	0.251	0.942

3.3

Graded care approach based on nursing risk assessment results

Based on the above nursing risk assessment results for the elderly population, this paper introduces a nursing care program that is more specific to the high nursing risk factors, and the following is a test of the hierarchical nursing care approach designed in this paper based on the nursing risk assessment results.

3.3.1

Experimental subjects and methods

This study was conducted on 150 cases of elderly patients admitted to a nursing home from August 2023 to August 2024. They included 61 cases of diabetic patients, 69 cases of hypertensive patients, and 20 cases of other cardiovascular disease patients. They were randomly divided into two groups, 75 cases of which adopted conventional nursing care as the control group, with 38 male cases and 37 female cases, aged 55-78 years old, with an average of (62.1±5.5) years old. The other 75 cases took nursing risk assessment and graded care as an experimental group, of which 36 cases were male and 39 cases were female, aged 54-73 years old, with an average of (61.2±6.1) years old. There was no significant difference in the health status of the subjects in the two experimental groups (P>0.05), and they were comparable.

3.3.2

Nursing risk assessment experiment

The control group was given routine nursing care, i.e., the corresponding nursing measures corresponding to the disease. The experimental group added nursing risk assessment and graded nursing care on the basis of routine care, and the specific measures were as follows.

CGRA nursing risk assessment form was adopted to assess nursing risk from 6 entries, such as nursing risk history, medical diagnosis, walking assistance, intravenous infusion/intubation or use of special drugs, patient’s gait and patient’s cognition. A total score of 125 points was assigned and graded according to the assessment results, with <24 being low risk, 25-45 being moderate risk, and >45 being high risk. Patients were assessed once a week during hospitalization, and targeted risk prevention nursing measures were implemented according to the grading results, namely: 1)

Introduce the relevant hospital and ward environments to low nursing risk level patients and their families. Conducting patient’s knowledge of preventive nursing risks, such as how to get in and out of bed correctly, how to use walking aids, and the use of utilities. Teach family members to help patients clean up factors that may cause nursing risks, such as obstacles and slippery floors.

2)

In addition to adopting nursing measures for patients and their families with medium nursing risk level, family members should be instructed to pay attention to the patient’s mobility and notify the nursing staff in time if any abnormality is found. The signboard of “medium nursing risk” is hung in the ward and bed, and the nursing staff carries out nursing risk checking every 8h for the patients with signboards.

3)

In addition to high-risk nursing measures for patients with high nursing risk level, “high nursing risk” signboards should be hung in the wards and beds, nursing staff should conduct nursing risk check every 4h, and family members should closely monitor the patient’s behavior, and stay with the patient when the patient is going to the toilet or taking a bath without any special circumstances, or inform the nursing staff if the patient leaves the room for a longer period of time. Nursing staff should intensify their rounds. Appropriately reduce the time and scope of the patient’s activities and avoid going out of the ward alone. Nursing staff should focus on explaining the situation of high-risk patients during shift handover.

SPSS 18.0 software was used for analysis, and Table 6 shows the comparison of nursing risk rate and nursing satisfaction between the two groups. Nursing risk occurred in 8 patients in the control group without obvious injuries, and the nursing risk rate was 10.67%, and no nursing risk occurred in the experimental group, and the difference between the two groups was statistically significant (P<0.05). The nursing satisfaction of the experimental group was 100%, which was significantly higher than that of the control group, and the difference was statistically significant (P<0.05).

Table 6.

The comparison of nursing risk incidence and nursing satisfaction

Group	Risk incidence (%)	Satisfaction evaluation
Group	Risk incidence (%)	Unsatisfied	Good	Satisfied
Control group	8 (10.67)	2 (2.7)	31 (41.3)	42 (56)
Experimental group	0 (0)	0 (0)	16 (21.3)	59 (78.7)
χ²	4.138	1.022	6.978	7.698
P	0.031	0.135	0.004	0.005

Table 7 shows the comparison of nursing risk assessment between admission and discharge in the experimental group. There was no patient with high nursing risk level in the experimental group at the time of discharge, and there was only one patient with medium nursing risk level, and the nursing risk assessment score was significantly lower than that at the time of admission, and the difference was statistically significant (P<0.05). It can be seen that after considering the factors causing nursing risk and carrying out targeted nursing care optimization, the nursing care for the elderly has been significantly improved, which also proves the practical significance of evaluating the nursing risk of the elderly through the logistics regression model.

Table 7.

The comparison of nursing risk incidence assessment of experimental group

Time	Low risk	Medium risk	High risk	Assessment of nursing risk
Admission	42 (56)	18 (24)	15 (20)	39.85±4.16
Discharge	74 (98.7)	1 (1.3)	0 (0)	19.32±3.11
χ²/t	32.358	14.873	15.981	26.872
P	0.000	0.000	0.000	0.000

4

Conclusion

This paper focuses on the use of Logistic regression modeling and CGRA Geriatric Nursing Risk Scale to assess the factors that predispose to risk in the care of the geriatric population and the use of a hierarchical care approach to develop a more targeted geriatric nursing care service. In the sample taken, the incidence of nursing risk was 46.2%, and the incidence of primary and secondary nursing risk was 27.2% and 19%, respectively. It was also found that the incidence of nursing risk was higher in females than in males (p<0.05), and the incidence of primary and secondary nursing risk in the elderly increased with age. Factors such as fixed armrest balance and chronic diseases were risk factors for elderly care, while increased physical activity was a protective factor. Graded nursing care was performed according to the nursing risk assessment results, and the experimental results showed that there were no risk events in the experimental group using the graded nursing care method, and the satisfaction of the elderly was 100%, and the nursing risk assessment score was significantly reduced (P<0.05). It can be seen that through nursing data and regression analysis, caregivers can optimize elderly care services more accurately based on nursing risk assessment.

Language:: English

Publication timeframe:: 1 times per year
Journal Subjects:: Life Sciences, Life Sciences, other, Mathematics, Applied Mathematics, General Mathematics, Physics, Physics, other

Journal RSS Feed

Nursing Risk Assessment for the Elderly and Its Data-Driven Optimization Model

Jinping Li

Xiaoyuan Hu

Published Online: Sep 26, 2025

Received: Jan 26, 2025

Accepted: May 07, 2025

DOI: https://doi.org/10.2478/amns-2025-1069

KeywordsLogistic regression, CGRA, Nursing risk assessment, Elderly group, Assessment optimization model

© 2025 Jinping Li et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Keywords
Logistic regression, CGRA, Nursing risk assessment, Elderly group, Assessment optimization model