Health and Social Care in the Community (2014) 22(5), 488–496
doi: 10.1111/hsc.12108
The risk factors for hospital re-admission in medical patients in Singapore Moon Fai Chan
PhD CStat
1
and Frances K. Y. Wong
RN PhD FAAN FHKAN
2,3
1
Alice Lee Centre for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Faculty of Health and Social Sciences, Hong Kong Polytechnic University, Hong Kong SAR, China and 3School of Nursing, Hong Kong Polytechnic University, Hong Kong SAR, China
2
Accepted for publication 10 January 2014
Correspondence Dr M. F. Chan Alice Lee Centre for Nursing Studies Yong Loo Lin School of Medicine National University of Singapore Singapore 117597 E-mail: [email protected]
What is known about this topic
•
•
Unplanned re-admissions are associated with age, gender, poor functional ability, lack of support, previous length of stay and type of disease. Most studies consider limited risk factors or fail to take confounding into account.
What this paper adds
•
• •
Unplanned re-admission was associated with being unemployed, having chronic obstructive pulmonary disease or having abnormal respiratory patterns; not being re-admitted was associated with doing regular daily activities, being assisted by a social worker, being referred to other health professionals when sick, or receiving a health education programme before discharge during a previous admission. Patients’ subjective enabling and need outcomes have an effect on re-admission. More effective educational care programmes incorporating patients’ own subjective health assessments are needed before discharge.
488
Abstract Unplanned hospital re-admissions are common, expensive and often unpreventable in the community. The study aimed to identify risk factors associated with unplanned hospital re-admission in Singapore. In a crosssectional survey, 1509 patients admitted to the medical wards of a large acute hospital in Singapore during 2010 were recruited (78.8% response rate), data being collected using a structured questionnaire based on the Andersen behavioural model underlying healthcare use. The dependent variable was re-admission within 28 days, with independent variables in the four areas of predisposing characteristics, needs, enabling resources and health behaviour. Hierarchical logistic regression was used to evaluate the risk factors associated with unplanned hospital re-admission. There were 222 inpatients re-admitted (14.7%) within 28 days and the final model showed that patients who were unemployed (OR = 1.5; 95% CI = 1.1–2.1) and had chronic obstructive pulmonary disease (OR = 2.0; 95% CI = 1.1–3.7) with abnormal respiratory patterns (OR = 1.6; 95% CI = 1.1–2.2) were more likely to be re-admitted. Less likely to be re-admitted were patients doing regular daily activities (OR = 0.7; 95% CI = 0.5–0.9), those assisted by a social worker (OR = 0.3; 95% CI = 0.2– 0.6), those referred to other health professionals when sick (OR = 0.6; 95% CI = 0.4–0.7) and those who had received health education programmes before discharge in the previous admission (OR = 0.7; 95% CI = 0.4–0.9). Unplanned re-admissions are a concern to healthcare providers because this suggests that patients are discharged with unresolved problems that reflect ineffective care in hospital. This study provides evidence to prompt more effective discharge educational care programmes that incorporate patients’ enabling and need outcomes, thereby reducing re-admission rates. Community-based healthcare should play an important role in reducing patients’ re-admission rates. Keywords: nursing, Singapore, unplanned re-admission
Introduction In Singapore, the cost of unplanned hospitalisation accounts for a large part of the high healthcare expenditure, estimated to be 35% of the total direct cost of medical care (Cao et al. 2006). In the United States, approximately 25% of hospitalised Medicare fee-for-service beneficiaries are admitted to the hospital within 30 days of discharge, accounting for US $17.6 billion in federal expenditures and most of these admissions are often unplanned re-admission (Naylor et al. 1999, 2004). The rate for unplanned re-admission can vary from country to country, ranging from 5% to 17% (Garcia et al. 1998, Kossovsky et al. 1999, Au et al. 2002, Wong et al. 2002, Allaudeen et al. 2011). © 2014 John Wiley & Sons Ltd
Unplanned re-admission
To reduce pressure and costs in the hospital system, community-based services for older people have evolved as one method of preventing unplanned re-admission. Findings from previous studies have shown that unplanned re-admissions were associated with factors like age (Kossovsky et al. 1999), ethnicity (Allaudeen et al. 2011), gender (Chan et al. 2008), poor functional ability (Culler et al. 1998), lack of financial support (Wong et al. 2010), poor social and family support (Andersen & Davidson 2001, Wong et al. 2004), previous length of stay (Kwok et al. 2004), having chronic obstructive pulmonary disease (COPD) patients (Cao et al. 2006) and inadequate discharge planning (Naylor et al. 2004). Most studies considered a limited number of specific risk factors in a single study and some did not fully take into account potential confounding among variables. Therefore, it is not surprising that all these studies had yielded inconsistent results and many aspects of this phenomenon are still poorly understood. In reviewing the literature, it is found that Andersen offers a theoretical model that may help give a better explanation of the phenomenon of hospital re-admissions (Andersen & Davidson 2001). Andersen’s model uses a systems approach to explain healthcare utilisation behaviour, and links health utilisation to outcomes. The model suggests that there are four main decision components that may predict the use of health services: predisposing factors; needs; enabling resources; and health behaviours. Predisposing factors include age, sex, marital status, education, ethnicity and occupation. Need factors refer to the most immediate cause of health service use, from functional and health problems that generate the need for healthcare services. Examples are individuals’ perceived and evaluated functional capacity, symptoms and general state of health. Enabling factors encompass family and community resources and accessibility of those resources, such as the means and know-how to access health services, available health personnel and facilities. Health behaviour factors refer to an individual’s attitude towards health services, knowledge about disease and values. By using the model’s relationships, a researcher can determine the directionality of the effect following a change in an individual’s characteristics or environment (Andersen & Davidson 2001). Aim In an ever-changing environment, the aim of this study was to assess which health service usage and other factors best predict unplanned hospital re-admissions in Singapore. © 2014 John Wiley & Sons Ltd
Methods Design and samples A cross-sectional survey was conducted during 2010 in the medical wards of a large acute hospital in Singapore. All patients who were admitted during the study period were included. Those who were admitted for scheduled procedures, were transferred from another hospital, or died within 72 hours of admission were excluded. Data were collected by using a structured questionnaire undertaken within 3 days of participants’ admission. Participants’ clinical and utilisation data were collected via the hospital record system. The data were collected by trained research assistants. A pilot study (n = 50) was conducted first to assess the logistics required for the main study. In the main study, the research assistant was stationed in the ward and those patients who were interested in the study were referred to the research assistant who then explained the study in detail. If the patient agreed to participate in the study and gave written consent, a face-to-face interview took place immediately. It took approximately 30 minutes to complete the questionnaire. The sample size calculation was based on the readmission rate within 28 days in previous studies (Kossovsky et al. 1999, Naylor et al. 1999, Cao et al. 2006, Chan et al. 2008, Wong et al. 2010), which suggested that it was around 11.2%. The number of participants recruited allowed this to be estimated with a precision of 1.6% (nQuery Advisor 2000). Instrument The Andersen model was used as a guiding model in our selection of variables (Andersen & Davidson 2001). This model was selected because it is based on data in four relevant areas – predisposing factors, need factors, enabling resources and health behaviours – that have been used by many investigators to explore unplanned admission (Wong et al. 2004, Chan et al. 2008). The instrument used in this study consisted of four parts: (i) predisposing factors, e.g. age, ethnicity, gender, education, occupation, having Medishield and/or Medifund, the disease category, number of re-admissions and previous admission and length of stay of previous admission (days); (ii) enabling resources factors, e.g. patients’ financial status, and social supports, such as social worker services; (iii) need factors, e.g. comprehensive medical assistance/ insurance, such as need for hospital admission, treatment needs, expressed needs of different healthcare services and whether they received post-discharge health education; and (iv) health behaviour factors, 489
M. F. Chan & F. K. Y. Wong
e.g. personal health practices, health outcomes, e.g. activities of daily living (ADL) scores (eight items ranging from 0 to 16, the higher the score, the more dependent), self-reported quality of life (QoL) (one item, from 1 to 5, the higher the score, the more unsatisfactory) (Zullig et al. 2006). All clinical health data, such as vital signs, respiratory patterns, and ability to communicate, were retrieved from the hospital record system. Ethical considerations Ethical approval was granted by the ethics committee of the target hospital under the National Healthcare Group (DSRB Ref: D/09/196) and all participants were asked to sign a consent form before they completed the questionnaire. Participants were told that they had the right to withdraw at any time for any reason, all information would be treated confidentially and they could complete the questionnaire in their own time, before handing it back to the research assistant. Data analysis Descriptive statistics were used to summarise participants’ characteristics by re-admission status. Readmission was defined as being admitted again within 28 days. The period used for this study was the same as another study conducted by the research team in Hong Kong (Wong et al. 2005). Chi-square tests, Fisher’s exact test and t-tests were used to evaluate the association of predisposing, enabling, need and health variables with re-admission status. Then variables with a P < 0.05 in the bivariate analysis were included in the multiple hierarchical logistic regression analyses, where re-admission status was the dependent variable. Re-admission was regressed on the following sequence of models: predisposing factors only (model 1); predisposing and enabling factors (model 2); predisposing, enabling and need factors (model 3); predisposing, enabling, need and health factors (model 4). The odds ratio (OR) and 95% confidence interval (CI) were estimated to identify the odds of the contributing factors in predicting re-admission. Data were analysed using IBM SPSS 19 (Chicago, NY, USA) with a statistical significance level of a = 0.05.
Results Sample characteristics In total, 1509 inpatients (response rate = 78.4%) were interviewed during the study period. On average, the patients had a mean age of 60.8 [standard deviation (SD) = 16.5] ranging from 21 to 96 (median = 62.0), 490
and 222 were re-admitted within 28 days of discharge with an estimated re-admission rate of 14.7% (95% CI = 13.0–16.6). The participants consisted of 839 males (55.6%) and 670 females (44.4%). The majority ethnic group was Chinese (70.4%), and participants had primary/secondary education level (62.1%) and were financially dependent (90.3%). Most of the subjects were married (76.4%) with over 50% retired/housewives (n = 766) who were living with family members. The re-admitted group had a mean age of 61.1 (SD = 16.3), while the non-re-admitted group had a mean age of 60.8 (SD = 16.5; t = 0.239, df = 1507, P = 0.811). Re-admission within 28 days was not associated with previous admission: 74.3% of those re-admitted had previously been admitted compared to 69.0% of those who were not re-admitted (v2 = 2.30, df = 1, P = 0.129). For those with a previous hospital stay, the previous length of stay for the re-admitted group was 7.2 days (SD = 6.5) and the non-re-admitted group was 6.7 days (SD = 6.2; U = 137,436.00, P = 0.284). Statistically significant associations were found between re-admission status on gender (P = 0.049), occupation (P < 0.001), chronic diseases e.g. respiratory problems (P = 0.005), renal failure (P = 0.008), COPD (P < 0.001) and diabetes (P = 0.022) (see Table 1). In Table 2, among all patients, only 24.6% (n = 370) reported that they were the main breadwinner in their family and more than 50% felt that they earned sufficient money (n = 814). However, more patients in the re-admitted group than in the non-readmitted group needed social support (14.0% vs. 5.8%, P < 0.001) or a social worker (10.8% vs. 3.0%, P < 0.001). Regarding the needs items, more patients in the re-admitted group felt that it was necessary for someone to take care of them when they were sick (64.3% vs. 54.0%, P = 0.005) and more were referred to see other health professionals (64.9% vs. 51.8%, P < 0.001). Re-admitted patients were less likely to see a private doctor (25.7% vs. 36.3%, P = 0.002) but more likely to go to hospital (27.0% vs. 20.5%, P = 0.042) or A&E (16.2% vs. 9.5%, P = 0.004). More patients in the re-admitted group had received some form of education programme before discharge from their previous admission (68.5% vs. 58.1%, P = 0.004). The majority of all patients reported their ADL score as independent (range 16–20) and most of their health indicators were normal on their current admission. Almost half of them (n = 751) reported a very satisfactory/satisfactory perceived QoL score. However, there were significant differences between groups. Readmitted patients had poorer ADL scores (P = 0.003), less satisfactory perceived QoL (P = 0.012) and were more likely to have abnormal health indicators. © 2014 John Wiley & Sons Ltd
Unplanned re-admission
Table 1 Predisposing factors by re-admission status (N = 1509) Re-admission within 28 days
Factors Gender (male) Age (years) 21–49 50–69 70–96 Mean (SD) Median [range] Marital status Single Married Divorced/separated/widowed Ethnicity Chinese Malay Indian Others Education level No formal education Primary/secondary Tertiary and above Occupation Employed Unemployed Retired/housewife/others Previous admission No Yes Previous length of stay (days) Mean (SD) Median [range] Medisave (yes) Medical insurance (yes) Medishield (yes) Medifund (yes) Eldershield (yes) Chronic disease Respiratory (yes) Heart failure (yes) Renal failure (yes) COPD (yes) Hypertension (yes) Asthma (yes) Stroke (yes) Diabetes (yes)
Total
Yes (n = 222)
No (n = 1287)
v2 test
n (%)
n (%)
n (%)
Statistic
839 (55.6)
137 (61.7)
702 (54.5)
3.94
0.049*
340 661 508 60.8 62.0
49 103 70 61.1 61.6
291 558 438 60.8 62.2
0.78
0.679
0.24‡
0.811
(22.5) (43.8) (33.7) (16.5) [21.0–96.0]
(22.1) (46.4) (31.5) (16.3) [21.0–96.0]
(22.6) (43.4) (34.0) (16.5) [21.0–95.0]
P-value
231 (15.4) 1147 (76.4) 124 (8.3)
38 (17.2) 163 (73.8) 20 (9.0)
193 (15.1) 984 (76.8) 104 (8.1)
0.98
0.611
1062 206 151 90
161 35 17 9
901 171 134 81
(70.0) (13.3) (10.4) (6.3)
4.05
0.256
(70.4) (13.7) (10.0) (6.0)
(72.5) (15.8) (7.7) (4.1)
380 (25.4) 931 (62.1) 187 (12.5)
54 (24.4) 148 (67.0) 19 (8.6)
326 (25.5) 783 (61.3) 168 (13.2)
4.19
0.123
538 (35.7) 202 (13.4) 766 (50.9)
63 (28.5) 51 (23.1) 107 (48.4)
475 (37.0) 151 (11.8) 659 (51.3)
22.12
doi: 10.1111/hsc.12108
The risk factors for hospital re-admission in medical patients in Singapore Moon Fai Chan
PhD CStat
1
and Frances K. Y. Wong
RN PhD FAAN FHKAN
2,3
1
Alice Lee Centre for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Faculty of Health and Social Sciences, Hong Kong Polytechnic University, Hong Kong SAR, China and 3School of Nursing, Hong Kong Polytechnic University, Hong Kong SAR, China
2
Accepted for publication 10 January 2014
Correspondence Dr M. F. Chan Alice Lee Centre for Nursing Studies Yong Loo Lin School of Medicine National University of Singapore Singapore 117597 E-mail: [email protected]
What is known about this topic
•
•
Unplanned re-admissions are associated with age, gender, poor functional ability, lack of support, previous length of stay and type of disease. Most studies consider limited risk factors or fail to take confounding into account.
What this paper adds
•
• •
Unplanned re-admission was associated with being unemployed, having chronic obstructive pulmonary disease or having abnormal respiratory patterns; not being re-admitted was associated with doing regular daily activities, being assisted by a social worker, being referred to other health professionals when sick, or receiving a health education programme before discharge during a previous admission. Patients’ subjective enabling and need outcomes have an effect on re-admission. More effective educational care programmes incorporating patients’ own subjective health assessments are needed before discharge.
488
Abstract Unplanned hospital re-admissions are common, expensive and often unpreventable in the community. The study aimed to identify risk factors associated with unplanned hospital re-admission in Singapore. In a crosssectional survey, 1509 patients admitted to the medical wards of a large acute hospital in Singapore during 2010 were recruited (78.8% response rate), data being collected using a structured questionnaire based on the Andersen behavioural model underlying healthcare use. The dependent variable was re-admission within 28 days, with independent variables in the four areas of predisposing characteristics, needs, enabling resources and health behaviour. Hierarchical logistic regression was used to evaluate the risk factors associated with unplanned hospital re-admission. There were 222 inpatients re-admitted (14.7%) within 28 days and the final model showed that patients who were unemployed (OR = 1.5; 95% CI = 1.1–2.1) and had chronic obstructive pulmonary disease (OR = 2.0; 95% CI = 1.1–3.7) with abnormal respiratory patterns (OR = 1.6; 95% CI = 1.1–2.2) were more likely to be re-admitted. Less likely to be re-admitted were patients doing regular daily activities (OR = 0.7; 95% CI = 0.5–0.9), those assisted by a social worker (OR = 0.3; 95% CI = 0.2– 0.6), those referred to other health professionals when sick (OR = 0.6; 95% CI = 0.4–0.7) and those who had received health education programmes before discharge in the previous admission (OR = 0.7; 95% CI = 0.4–0.9). Unplanned re-admissions are a concern to healthcare providers because this suggests that patients are discharged with unresolved problems that reflect ineffective care in hospital. This study provides evidence to prompt more effective discharge educational care programmes that incorporate patients’ enabling and need outcomes, thereby reducing re-admission rates. Community-based healthcare should play an important role in reducing patients’ re-admission rates. Keywords: nursing, Singapore, unplanned re-admission
Introduction In Singapore, the cost of unplanned hospitalisation accounts for a large part of the high healthcare expenditure, estimated to be 35% of the total direct cost of medical care (Cao et al. 2006). In the United States, approximately 25% of hospitalised Medicare fee-for-service beneficiaries are admitted to the hospital within 30 days of discharge, accounting for US $17.6 billion in federal expenditures and most of these admissions are often unplanned re-admission (Naylor et al. 1999, 2004). The rate for unplanned re-admission can vary from country to country, ranging from 5% to 17% (Garcia et al. 1998, Kossovsky et al. 1999, Au et al. 2002, Wong et al. 2002, Allaudeen et al. 2011). © 2014 John Wiley & Sons Ltd
Unplanned re-admission
To reduce pressure and costs in the hospital system, community-based services for older people have evolved as one method of preventing unplanned re-admission. Findings from previous studies have shown that unplanned re-admissions were associated with factors like age (Kossovsky et al. 1999), ethnicity (Allaudeen et al. 2011), gender (Chan et al. 2008), poor functional ability (Culler et al. 1998), lack of financial support (Wong et al. 2010), poor social and family support (Andersen & Davidson 2001, Wong et al. 2004), previous length of stay (Kwok et al. 2004), having chronic obstructive pulmonary disease (COPD) patients (Cao et al. 2006) and inadequate discharge planning (Naylor et al. 2004). Most studies considered a limited number of specific risk factors in a single study and some did not fully take into account potential confounding among variables. Therefore, it is not surprising that all these studies had yielded inconsistent results and many aspects of this phenomenon are still poorly understood. In reviewing the literature, it is found that Andersen offers a theoretical model that may help give a better explanation of the phenomenon of hospital re-admissions (Andersen & Davidson 2001). Andersen’s model uses a systems approach to explain healthcare utilisation behaviour, and links health utilisation to outcomes. The model suggests that there are four main decision components that may predict the use of health services: predisposing factors; needs; enabling resources; and health behaviours. Predisposing factors include age, sex, marital status, education, ethnicity and occupation. Need factors refer to the most immediate cause of health service use, from functional and health problems that generate the need for healthcare services. Examples are individuals’ perceived and evaluated functional capacity, symptoms and general state of health. Enabling factors encompass family and community resources and accessibility of those resources, such as the means and know-how to access health services, available health personnel and facilities. Health behaviour factors refer to an individual’s attitude towards health services, knowledge about disease and values. By using the model’s relationships, a researcher can determine the directionality of the effect following a change in an individual’s characteristics or environment (Andersen & Davidson 2001). Aim In an ever-changing environment, the aim of this study was to assess which health service usage and other factors best predict unplanned hospital re-admissions in Singapore. © 2014 John Wiley & Sons Ltd
Methods Design and samples A cross-sectional survey was conducted during 2010 in the medical wards of a large acute hospital in Singapore. All patients who were admitted during the study period were included. Those who were admitted for scheduled procedures, were transferred from another hospital, or died within 72 hours of admission were excluded. Data were collected by using a structured questionnaire undertaken within 3 days of participants’ admission. Participants’ clinical and utilisation data were collected via the hospital record system. The data were collected by trained research assistants. A pilot study (n = 50) was conducted first to assess the logistics required for the main study. In the main study, the research assistant was stationed in the ward and those patients who were interested in the study were referred to the research assistant who then explained the study in detail. If the patient agreed to participate in the study and gave written consent, a face-to-face interview took place immediately. It took approximately 30 minutes to complete the questionnaire. The sample size calculation was based on the readmission rate within 28 days in previous studies (Kossovsky et al. 1999, Naylor et al. 1999, Cao et al. 2006, Chan et al. 2008, Wong et al. 2010), which suggested that it was around 11.2%. The number of participants recruited allowed this to be estimated with a precision of 1.6% (nQuery Advisor 2000). Instrument The Andersen model was used as a guiding model in our selection of variables (Andersen & Davidson 2001). This model was selected because it is based on data in four relevant areas – predisposing factors, need factors, enabling resources and health behaviours – that have been used by many investigators to explore unplanned admission (Wong et al. 2004, Chan et al. 2008). The instrument used in this study consisted of four parts: (i) predisposing factors, e.g. age, ethnicity, gender, education, occupation, having Medishield and/or Medifund, the disease category, number of re-admissions and previous admission and length of stay of previous admission (days); (ii) enabling resources factors, e.g. patients’ financial status, and social supports, such as social worker services; (iii) need factors, e.g. comprehensive medical assistance/ insurance, such as need for hospital admission, treatment needs, expressed needs of different healthcare services and whether they received post-discharge health education; and (iv) health behaviour factors, 489
M. F. Chan & F. K. Y. Wong
e.g. personal health practices, health outcomes, e.g. activities of daily living (ADL) scores (eight items ranging from 0 to 16, the higher the score, the more dependent), self-reported quality of life (QoL) (one item, from 1 to 5, the higher the score, the more unsatisfactory) (Zullig et al. 2006). All clinical health data, such as vital signs, respiratory patterns, and ability to communicate, were retrieved from the hospital record system. Ethical considerations Ethical approval was granted by the ethics committee of the target hospital under the National Healthcare Group (DSRB Ref: D/09/196) and all participants were asked to sign a consent form before they completed the questionnaire. Participants were told that they had the right to withdraw at any time for any reason, all information would be treated confidentially and they could complete the questionnaire in their own time, before handing it back to the research assistant. Data analysis Descriptive statistics were used to summarise participants’ characteristics by re-admission status. Readmission was defined as being admitted again within 28 days. The period used for this study was the same as another study conducted by the research team in Hong Kong (Wong et al. 2005). Chi-square tests, Fisher’s exact test and t-tests were used to evaluate the association of predisposing, enabling, need and health variables with re-admission status. Then variables with a P < 0.05 in the bivariate analysis were included in the multiple hierarchical logistic regression analyses, where re-admission status was the dependent variable. Re-admission was regressed on the following sequence of models: predisposing factors only (model 1); predisposing and enabling factors (model 2); predisposing, enabling and need factors (model 3); predisposing, enabling, need and health factors (model 4). The odds ratio (OR) and 95% confidence interval (CI) were estimated to identify the odds of the contributing factors in predicting re-admission. Data were analysed using IBM SPSS 19 (Chicago, NY, USA) with a statistical significance level of a = 0.05.
Results Sample characteristics In total, 1509 inpatients (response rate = 78.4%) were interviewed during the study period. On average, the patients had a mean age of 60.8 [standard deviation (SD) = 16.5] ranging from 21 to 96 (median = 62.0), 490
and 222 were re-admitted within 28 days of discharge with an estimated re-admission rate of 14.7% (95% CI = 13.0–16.6). The participants consisted of 839 males (55.6%) and 670 females (44.4%). The majority ethnic group was Chinese (70.4%), and participants had primary/secondary education level (62.1%) and were financially dependent (90.3%). Most of the subjects were married (76.4%) with over 50% retired/housewives (n = 766) who were living with family members. The re-admitted group had a mean age of 61.1 (SD = 16.3), while the non-re-admitted group had a mean age of 60.8 (SD = 16.5; t = 0.239, df = 1507, P = 0.811). Re-admission within 28 days was not associated with previous admission: 74.3% of those re-admitted had previously been admitted compared to 69.0% of those who were not re-admitted (v2 = 2.30, df = 1, P = 0.129). For those with a previous hospital stay, the previous length of stay for the re-admitted group was 7.2 days (SD = 6.5) and the non-re-admitted group was 6.7 days (SD = 6.2; U = 137,436.00, P = 0.284). Statistically significant associations were found between re-admission status on gender (P = 0.049), occupation (P < 0.001), chronic diseases e.g. respiratory problems (P = 0.005), renal failure (P = 0.008), COPD (P < 0.001) and diabetes (P = 0.022) (see Table 1). In Table 2, among all patients, only 24.6% (n = 370) reported that they were the main breadwinner in their family and more than 50% felt that they earned sufficient money (n = 814). However, more patients in the re-admitted group than in the non-readmitted group needed social support (14.0% vs. 5.8%, P < 0.001) or a social worker (10.8% vs. 3.0%, P < 0.001). Regarding the needs items, more patients in the re-admitted group felt that it was necessary for someone to take care of them when they were sick (64.3% vs. 54.0%, P = 0.005) and more were referred to see other health professionals (64.9% vs. 51.8%, P < 0.001). Re-admitted patients were less likely to see a private doctor (25.7% vs. 36.3%, P = 0.002) but more likely to go to hospital (27.0% vs. 20.5%, P = 0.042) or A&E (16.2% vs. 9.5%, P = 0.004). More patients in the re-admitted group had received some form of education programme before discharge from their previous admission (68.5% vs. 58.1%, P = 0.004). The majority of all patients reported their ADL score as independent (range 16–20) and most of their health indicators were normal on their current admission. Almost half of them (n = 751) reported a very satisfactory/satisfactory perceived QoL score. However, there were significant differences between groups. Readmitted patients had poorer ADL scores (P = 0.003), less satisfactory perceived QoL (P = 0.012) and were more likely to have abnormal health indicators. © 2014 John Wiley & Sons Ltd
Unplanned re-admission
Table 1 Predisposing factors by re-admission status (N = 1509) Re-admission within 28 days
Factors Gender (male) Age (years) 21–49 50–69 70–96 Mean (SD) Median [range] Marital status Single Married Divorced/separated/widowed Ethnicity Chinese Malay Indian Others Education level No formal education Primary/secondary Tertiary and above Occupation Employed Unemployed Retired/housewife/others Previous admission No Yes Previous length of stay (days) Mean (SD) Median [range] Medisave (yes) Medical insurance (yes) Medishield (yes) Medifund (yes) Eldershield (yes) Chronic disease Respiratory (yes) Heart failure (yes) Renal failure (yes) COPD (yes) Hypertension (yes) Asthma (yes) Stroke (yes) Diabetes (yes)
Total
Yes (n = 222)
No (n = 1287)
v2 test
n (%)
n (%)
n (%)
Statistic
839 (55.6)
137 (61.7)
702 (54.5)
3.94
0.049*
340 661 508 60.8 62.0
49 103 70 61.1 61.6
291 558 438 60.8 62.2
0.78
0.679
0.24‡
0.811
(22.5) (43.8) (33.7) (16.5) [21.0–96.0]
(22.1) (46.4) (31.5) (16.3) [21.0–96.0]
(22.6) (43.4) (34.0) (16.5) [21.0–95.0]
P-value
231 (15.4) 1147 (76.4) 124 (8.3)
38 (17.2) 163 (73.8) 20 (9.0)
193 (15.1) 984 (76.8) 104 (8.1)
0.98
0.611
1062 206 151 90
161 35 17 9
901 171 134 81
(70.0) (13.3) (10.4) (6.3)
4.05
0.256
(70.4) (13.7) (10.0) (6.0)
(72.5) (15.8) (7.7) (4.1)
380 (25.4) 931 (62.1) 187 (12.5)
54 (24.4) 148 (67.0) 19 (8.6)
326 (25.5) 783 (61.3) 168 (13.2)
4.19
0.123
538 (35.7) 202 (13.4) 766 (50.9)
63 (28.5) 51 (23.1) 107 (48.4)
475 (37.0) 151 (11.8) 659 (51.3)
22.12
