Rehabilitation of Older Chinese Patients With Different Cognitive Functions: How Do They Differ in Outcome?

Setting 

The study was performed at the Acute Geriatrics Unit of Grantham Hospital and the geriatrics unit of Tung Wah Hospital, Hong Kong Special Administrative Region, China. There are 38 and 12 geriatric beds in these 2 units, respectively.

Participants 

We recruited older patients (age ≥65y) admitted between May 2004 and August 2006. Patients with acute illnesses are admitted to the University Medical Unit, Queen Mary Hospital, for management first. After stabilization, they are transferred to either of the above 2 geriatric units for rehabilitation. Both units also accept elderly patients directly admitted from the emergency department and nursing homes.

Intervention 

Both units consist of a multidisciplinary team led by a geriatrics specialist. Comprehensive geriatric assessment is performed on all newly admitted patients. After assessment, an individualized care plan is formulated and all patients routinely receive daily physiotherapy and occupational therapy unless they are considered to be unsuitable. If needed, dieticians, speech therapists, and medical social workers are referred for assessment and treatment. Patient progress is discussed in a weekly multidisciplinary case conference where the progress and rehabilitation goals are reviewed. Patients are discharged from the hospital if the multidisciplinary team agrees the patient's progress has reached a plateau and the patient is safe to go home. If community rehabilitation is considered necessary, we refer patients to Geriatric Day Hospital for further training after discharged from the hospital.

Design 

This was a retrospective cohort study, and the study protocol was approved by the local institutional review board. We established a comprehensive database for all elderly patients in May 2004 to capture both independent and dependent variables.

Independent Variables 

Collected data included demography, LOS in the hospital, occurrence of pressure ulcers and urinary incontinence, number of diagnoses, and chief diagnoses leading to admission. Body mass index was measured by the nurses. A C-MMSE was performed by the occupational therapists within 48 hours of admission. C-MMSE is a Cantonese translation of the original MMSE. A local study showed that when using Diagnostic and Statistical Manual of Mental Disorder—Third Edition—Revised criteria for diagnosis of dementia, a cutoff score of 19/20 yielded a high sensitivity rate of 97.5% and specificity of 97.3%.29 If patients were unable to perform C-MMSE because of various reasons such as an impaired conscious state, medical instability, inability to respond to questions, severe deafness, dysphasia, language barrier, or refused cooperation, they were recorded as failed, a C-MMSE score was not given, and they were excluded from the final analysis.

Dependent Variables 

Our physiotherapists and occupational therapists measured the EMS (maximum score, 20) and Barthel Index (maximum score, 100) in each patient, respectively.30, 31 For patients who did not complete the rehabilitation program because of death or transfer to other hospitals, their last Barthel Index or EMS was used. We also analyzed the efficacy, relative efficacy, and efficiency in this study to judge rehabilitation outcome.

Statistical Analyses 

Stratification of C-MMSE Groups 

We stratified the patients according to their C-MMSE scores into 3 groups: less than 10, 10 to 19, and 20 or higher. The cutoff point at 20 or higher was chosen because the usual cutoff point for C-MMSE defining impaired cognitive function was less than 20. This was based on the original local validation study showing that a cutoff score of 19/20 yielded a high sensitivity and specificity.29 Demographic and clinical characteristics of the patients were examined in these 3 stratified age groups. In addition, we analyzed the Barthel Index and EMS on admission and discharge as well as the absolute and relative change of these scores across the 3 C-MMSE groups.

Univariate Analysis 

SPSSa was used in statistical analysis. Continuous variables were expressed as mean ± SD. We used the paired t test to compare the change of continuous variables within group. Chi-square tests and Fisher exact tests were used to compare categoric variables. Simple correlation was performed with the Pearson correlation test. We used the 1-way ANOVA to compare continuous variables among 3 different age groups. Bonferroni adjustment was used for multiple comparisons. Statistical significance was inferred by a 2-tailed P value of .05 or less.

Multivariate Analysis 

A satisfactory functional gain was defined as a Barthel Index MRFS efficacy and EMS MRFS efficacy of .25 or higher. We used the cutoff point of .25 because it represented 25% or more of each patient's possible potential for improvement in the Barthel Index and EMS. Multivariate analysis was performed by using stepwise logistic regression to delineate the independent predictive factors for a satisfactory functional gain after rehabilitation. Variables that had a significant association with satisfactory functional gain to a value of P less than .10 on univariate analyses were entered into the final multivariate analysis. ORs and 95% CIs were used to estimate the association between the independent factors of satisfactory functional gain. Statistical significance was inferred by a 2-tailed P value of .05 or less.

The Correlation Between Cognitive Status and Different Outcomes 

Table 2 summarizes the results of the Pearson correlations. A significant positive correlation was observed between admission C-MMSE discharge and total Barthel Index scores and EMS (both P<.001). No significant correlation was found between admission C-MMSE scores and Barthel Index efficacy (P=.48) as well as EMS efficacy (P=.39). However, the Barthel Index MRFS efficacy and efficiency, as well as EMS MRFS efficacy and efficiency, had significant correlations with admission C-MMSE (P=.001, P<.001, respectively).

Table 2. Correlation Between C-MMSE and Different Outcome Measures

	Measure	Admission C-MMSE
		Pearson Correlation	P
	Discharge Barthel Index	.512	<.001
	Barthel Index efficacy	.25	.48
	Barthel Index MRFS efficacy	.118	.001
	Barthel Index MRFS efficiency	.124	.001
	Discharge EMS	.49	<.001
	EMS efficacy	.031	.39
	EMS MRFS efficacy	.157	<.001
	EMS MRFS efficiency	.126	<.001

Functional and Motor Outcomes in the 3 Cognitive Groups 

Table 3 shows the total Barthel Index and EMS on admission and at discharge as well as the change during rehabilitation. Among patients of different C-MMSE groups, there were significant differences in the total Barthel Index and EMS on admission and at discharge, with lower discharge scores in low C-MMSE groups (P<.001). The difference in the change of the Barthel Index (Barthel Index efficacy) and EMS (EMS efficacy) during rehabilitation by 1-way ANOVA test was not significant among the 3 C-MMSE groups (Barthel Index efficacy, P=.30; EMS efficacy, P=.59). However, significant differences were observed in Barthel Index MRFS efficacy (P=.001), Barthel Index MRFS efficiency (P<.001), EMS MRFS efficacy (P<.001), and EMS MRFS efficiency (P=.001) in the 3 cognitive function groups, with the groups with better cognitive function having better results.

Table 3. Total Barthel Index and EMS Scores on Admission and at Discharge as Well as the Change of Barthel Index and EMS Scores During Rehabilitation in Different C-MMSE Groups

	Measure	All Patients (N=778)	C-MMSE Score <10 (n=146)	C-MMSE Score 10–19 (n=378)	C-MMSE Score ≥20 (n=254)	P (1-way ANOVA)
	Barthel Index admission	66.1±24.8	46.2±26.9	64.9±22.5	79.2±17.8	<.001†
	Barthel Index discharge	73.2±24.5⁎	51.9±29.4⁎	72.3±22.3⁎	86.5±12.3⁎	<.001†
	Barthel Index efficacy	7.08±11.3	5.76±13.5	7.42±11.5	7.33±9.31	.30
	Barthel Index MRFS efficacy	0.12±0.16	0.077±0.19	0.12±0.16	0.14±0.14	.001†
	Barthel Index MRFS efficiency	0.007±0.012	0.005±0.013	0.007±0.01	0.009±0.012	<.001†
	EMS admission	9.37±6.57	4.48±4.94	8.9±5.9	12.9±6.3	<.001†
	EMS discharge	11.5±6.6⁎	6.4±6.14⁎	11.1±6.0⁎	15.2±5.35⁎	<.001†
	EMS efficacy	2.17±3.35	1.9±3.87	2.2±3.25	2.27±3.17	.59
	EMS MRFS efficacy	0.21±0.31	0.13±0.25	0.2±0.3	0.27±0.33	<.001†
	EMS MRFS efficiency	0.014±0.03	0.008±0.03	0.014±0.027	0.019±0.03	.001†

NOTE. Values are mean ± SD.

⁎ Significant difference comparing Barthel Index or EMS discharge and Barthel Index or EMS admission within each C-MMSE group by using paired t test (P=.000). † Significant difference using 1-way ANOVA.

Independent Predictors 

One hundred fifty-three (19.7%) patients achieved a Barthel Index MRFS efficacy of .25 or higher, whereas 225 (28.9%) reached an EMS MRFS efficacy of .25 or higher. Multivariate analysis of variables associated with a satisfactory Barthel Index outcome (Barthel Index MRFS efficacy ≥.25) after rehabilitation included significant variables obtained from univariate analysis including hospital LOS, age, education, diagnosis of heart failure, infection other than pneumonia, and musculoskeletal problems (included osteoarthritis, gouty arthritis, low back pain secondary to degenerative causes). The LOS (OR=1.02; 95% CI, 1.01–1.03; P=.002) and diagnosis of musculoskeletal problems (OR=2.24; 95% CI, 1.25–4.00; P=.007) were both positive independent predictors for a satisfactory Barthel Index MRFS outcome (table 4). The significant variables after univariate analysis for satisfactory EMS outcome (EMS MRFS efficacy, ≥.25) were age, LOS, accommodation, admission C-MMSE, BMI, pressure ulcer, urinary incontinence, and diagnosis of musculoskeletal disorders. Multivariate analysis showed that LOS (OR=1.02; 95% CI, 1.00–1.03; P=.003), admission C-MMSE (OR=1.04; 95% CI, 1.02–1.07; P=.001), BMI (OR=1.05; 95% CI, 1.01–1.09; P=.006), and diagnosis of musculoskeletal disorder (OR=1.75; 95% CI, 1.01–3.00; P=.04) were significant positive independent predictors for EMS MRFS efficacy greater than or equal to .25, whereas urinary incontinence on admission was a significant negative predictor (OR=.69; 95% CI, .48–.98; P=.04) (see table 4).

Tables 4. Results of the Multivariate Analysis

	Measure	Barthel Index MRFS Efficacy ≥.25			EMS MRFS Efficacy ≥.25
		OR	CI	P	OR	CI	P
	LOS	1.02	1.01–1.03	.000	1.02	1.00–1.03	.003
	Musculoskeletal disorder	2.24	1.25–4.00	.007	1.75	1.01–3.00	.04
	C-MMSE	NA	NA	NA	1.04	1.02–1.07	.001
	BMI	NA	NA	NA	1.05	1.01–1.09	.006
	Urinary incontinence	NA	NA	NA	0.69	0.48–0.98	.04

Abbreviation: NA, not applicable.

Study Limitations 

There are certain limitations in our study. One important question is whether the results are generalizable. This study looks at Hong Kong residents who are predominantly Chinese, and the finding may not be applicable to other cultures. Being a retrospective study, data accuracy is reliant on the quality of documentation in the case notes and the entry into the computer database. Nevertheless, we believe that the large sample size in this study can to a certain extent minimize this potential flaw. Two hundred twenty-three of 1001 cases were excluded from this study because they failed in the C-MMSE assessment in the first 48 hours. We did not investigate the reasons behind each case who failed C-MMSE assessment. We would like to emphasize that failure to perform C-MMSE should not be equivalent to poor rehabilitation potential. Some cases failed simply because of severe deafness, expressive dysphasia, or language barriers. They could still participate in rehabilitation if the therapist knew how to make use of nonverbal instructions. Some patients might fail the C-MMSE assessment because of medical instability in the first 2 days. They could still benefit from rehabilitation after their acute medical problems were resolved. We only measured the functional and motor outcomes during inpatient treatment here. Further study is needed to examine the effect of cognitive status in maintaining the functional and motor gain after discharge from the hospital.