a b s t r a c t

Objective: Medical Alert Protection Systems (MAPS) are a form of assistive technology designed to support inde- pendent living in the care of elderly patients in the community. We aimed to investigate the utility of using such a device (eAlert! System) in elderly patients presenting to an Emergency Department (ED).

Methods: Elderly patients presenting to an ED were randomized to receive MAPS or telephone follow-up only (control arm). All patients were followed up at one-week, one-month and six-month post-intervention. A confi- dence scale (at 1 week, 1 month and 6 months) and EQ-5D score (at 6 months) were also administered.

Results: 106 and 91 participants enrolled in the MAPS and control arms respectively. Within both individual arms, there were significant reductions in the median number of ED visits and median number of admissions in the six month periods before, compared to after intervention (p b 0.01 for both). However, the reductions were not sig- nificantly different between the two arms. Among participants who have had one or more admissions during the six months period post intervention, the MAPS arm had significantly lower median total length of stay (8 days, Interquartile Range [IQR] = (4, 14)) compared to the control arm (15 days, IQR = (3, 25), p = 0.045). The median health state score for health state was significantly higher in the MAPS arm (70 IQR = (60,80) versus 60 IQR = (50,70), p = 0.008).

Conclusion: In this population of elderly ED patients, the use of a MAPS decreased length of stay for admissions and improved quality of life measures.

(C) 2017

Introduction

Singapore has a rapidly ageing population [1], a trend also seen in the developed world [2], but particularly worse in Asia [3]. Not only that, the old-age support ratio (ratio of residents aged 20-64 to resi- dents aged 65 years and above) is also declining [1]. These demographic changes exert pressures on healthcare infrastructure [3], including emergency care, as the elderly not only have higher prevalence of age- related chronic diseases, but present more frequently to Emergency

* Corresponding author at: Department of Emergency Medicine, Singapore General Hospital, Outram Road, Singapore 169608, Singapore.

E-mail address: [email protected] (M.E.H. ONG).

Departments (EDs) with acute illnesses and exacerbations of chronic ill- nesses, often atypically, and with worse outcomes [4-6].

Indeed, geriatric ED attendance has been increasing in Singapore [7], and is observed to contribute to greater resource utilization [8] and fre- quent attenders [9]. Not only does this increased load in itself impose a strain on ED resources, the decreasing old-age support ratio means that it is becoming increasingly difficult to discharge medically stable elderly patients to reliable caregivers, as many elderly either stay alone or are alone at home when their caregivers are out for work. In a national sur- vey, 7.5% of Singapore residents aged 55 and above reported staying alone [10]. It is sometimes necessary for clinicians to have a lower threshold for admission of these patients for observation of atypical symptoms or for nursing support. These admissions can be associated

https://doi.org/10.1016/j.ajem.2017.09.027

0735-6757/(C) 2017

with hospitalisation related complications [11], and increase consump- tion of scarce bed resources in settings of ED overcrowding and inpa- tient bed shortage [12].

These are some factors driving the use of assistive technology in the care of the elderly. Medical Alert Protection Systems (MAPS) are a form of assistive technology designed to support independent living in peo- ple by enabling them to gain fast assistance in an emergency. Typically the person accesses the emergency service by pressing a button on a portable device which is connected to a base unit in the house and com- municates with a 24-hour call center. These have been employed in var- ious Community settings, with differing outcomes in terms of health utilization and allaying anxiety in the users. However there have been few controlled trials evaluating the efficacy of such devices, with only one previous randomized trial in the emergency setting [13-15]. The ef- ficacy of devices in these studies have been mixed, included postulated benefits of reduced anxiety about falling, reduced anxiety for the person’s family, increased confidence in performing daily activities and prolonging the time people can remain independent staying alone [16].

This study aimed to evaluate the eAlert! Device, in a randomized clinical trial in the Asian ED setting. We hypothesized that the use of the eAlert! System in the care of elderly patients at risk of falls discharged from the Emergency Department would lead to 1) reduction in subsequent ED revisit rate compared to baseline, 2) reduction in number of hospitalizations compared to baseline, and 3) reduction in total length of stay of admissions for patients.

Methods

Setting

Singapore is an urbanized, island city-state in South-east Asia with a population of 5.3 million and land area of only 715.1 km² [17]. At the time of the study, Singapore General Hospital was the oldest of the six public general hospitals [18], and had an annual ED census of 143,000 in the year 2014.

Study design

This study was a randomized controlled trial. The recruitment period was from December 2012 to September 2014. Participants were allocat- ed to receive either MAPS (MAPS arm) or telephone follow-up only (control arm), via a block randomization process. Randomization was initially in an allocation ratio of 1:1, and in April 2014, was changed to 3:1 (three intervention to every control) due to higher drop outs from the MAPS arm. Block sizes were not revealed to the clinical study team. Baseline data was collected during a six month pre-intervention period where the ED did not have a protocol to routinely monitor elder- ly patients post discharge, via retrospective review of electronic medical records.

Recruitment was performed during weekdays, on morning and eve- ning shifts (0800-2200 h), by two research nurses. ED physicians would notify research nurses of any potential candidate over 65 year age who had a fall. These candidates were then screened by the research nurses, including determining Inclusion/exclusion criteria and administering the Morse Fall Scale Risk Assessment. After informed consent was ob- tained, the research nurse would obtain an opaque envelop with the al- located study group designation. Allocation was made, revealed to and discussed with the patient before the ED Disposition decision. Due to practical difference in treatment, there was no blinding of participants. However, the data analyst was blinded to allocation.

All patients were followed up by telephone interview by the two re- search nurses at one week, one month and six months post discharge. In addition, a quality of life assessment was conducted at the six months mark (see “Variables and outcome measures”). For patients who were receiving inpatient care during these milestones, they were followed

up by a review of the ED and inpatient electronic medical records post discharge.

Study population

All patients aged 65 years and above were eligible for enrolment if they 1) lived alone, were alone most of the day, or lived with someone who was unable to get to the phone in an emergency, and 2) had expe- rienced a fall in the preceding six months that required medical atten- tion or had at least low-to-moderate risk of fall as assessed by the Morse Fall Scale Risk Assessment Chart.

The exclusion criteria were 1) insufficient physical and Cognitive function to wear and operate the personal alert system, 2) unwilling to wear the device 24 h a day and to activate the system if necessary, and 3) inability of user or caregiver to replace batteries in response to a voice prompt.

Intervention

In the MAPS arm, the patients were allocated to a Medical Alert Pro- tection System in the form of the eAlert! System (Fig. 1). eAlert! (Active Medical [Asia Pacific] Pte Ltd., Singapore) is a personal emergency re- sponse system with around-the-clock call center monitoring support service. The devices consist of a pendant and a base unit. The pendant weighs 113 g, is water resistant and would be worn by the patient at all times around the neck, even into the shower. The pendant allows for three-way voice communication with the call center and a registered nurse anywhere within a 180 m radius from the base unit. In the case of an emergency, the patient can activate the call center with a single but- ton press. The call center has direct access to a nurse during office hours who can provide medical advice guided by protocols. Where needed, the call center can alert the national ambulance service and next-of- kin. The call center operator can also provide the patient’s medical re- cords (a summary of which are kept on file) to medical personnel if needed. It will emit a “beep” if it is out of range from the base unit. The vendor provided home installation services and training to the patient’s family if needed, and thereafter responded to technical diffi- culties with the devices.

In the control arm, patients received only telephone follow-up, i.e. at the one week, one month and six months. The same follow up questions regarding self-reported confidence and quality of life were used in the interview, similar to the intervention group.

Variables and outcome measures

The primary outcomes were 1) reduction in ED visit rate compared to baseline, 2) reduction in number of hospitalizations compared to baseline, and 3) total length of stay for admitted patients. Outcomes

(1) and (2) were also collected for a six month pre-intervention period as baseline data.

The secondary outcomes were 1) self-reported confidence (on a ten- point Likert scale) at 1 week, 1 month and 6 months post-intervention,

2) quality of life (EQ-5D) questionnaire at six months post-intervention, which included a self-reported health status on a 0-100 scale, and 3) mortality at six months. Mortality data was obtained by review of hos- pital electronic records and also verified by caregiver interview as needed.

The EQ-5D [19] is a standardized non-disease-specific instrument for valuing health-related quality of life, and its regionalized translations for non-English speaking Singaporean patients have been validated [20-23]. The self-report confidence scale consisted of a single question administered over the phone in the form of “How confident are you that you can get timely help during emergency at home?” and scored on a ten-point Likert scale. Other data collected included demographic data, comorbidities, social history including family situation, chief com- plaint of the ED visit and ED diagnosis.

Fig. 1. Study population flow diagram.

Sample size calculation

A sample size of 73 per arm was originally calculated to achieve 80% power to detect a difference of 20% in initial admission rate between the 2 groups (90% in control and 70% in intervention) at a significance level (alpha) of 0.05 using a two-sided z-test with continuity correction. These results assumed that 2 sequential tests are made using the O’Brien-Flem- ing [24] spending function to determine the test boundaries and patients would be recruited over 2 years. Sample size calculation was done using PASS 2008 [25]. However due to a higher dropout rate in the intervention arm, sampling rate had to be adjusted midway through the trial.

Statistical analysis

Categorical data such as gender and race were summarized as count and percentages. Scale or ordinal data such as number of hospitalizations were summarized as median, Interquartile Range (IQR) and range; the corresponding statistical comparison of the 2 types of data between the telephone follow-up and Medical Alert arms was performed using Chi- square test and Mann-Whitney test. Statistical analyses were performed using SPSS 19.0. The significance threshold was set at 0.05.

Ethics approval

Ethics approval for this study was approved by the Central Institu- tional Review Board at the Singapore General Hospital.

Results

Population flow

A total of 850 patients were screened according to the study criteria, 653 patients were excluded as they matched the exclusion criteria. Ma- jority of the patients were not keen to wear the device for 24 h a day thus was excluded.

The randomization procedure recruited 106 participants into the MAPS arm and 91 participants into the telephone follow-up arm (see Fig. 1). 34 participants from the MAPS arm and 1 participant from the telephone follow-up arm dropped out, with reasons given in Fig. 1.A re- sultant 72 participants from the MAPS arm and 90 participants from the telephone follow-up arm qualified for analysis.

Baseline characteristics

Table 1 shows the baseline characteristics of the subjects. All demo- graphic, comorbidity characteristics and presenting chief complaints were similar between the two arms.

Primary outcomes

In terms of decrease in the median number of ED visits in the six months post-intervention compared to the pre-intervention (baseline) phase, there was no significant difference between the two study arms (p = 0.88) (see Table 2). However, within both individual arms, there

Table 1

Baseline characteristics of study participants.

Characteristics	Telephone follow-up only	Medical Alert Protection System	p-Value
	(control) (N = 90)	(interventional) (N = 72)
Gender - no. (%)			0.744
Male	32 (35.6%)	28 (38.9%)
Female	58 (64.4%)	44 (61.1%)
Age group - no. (%)			0.579
65-75 years	42 (46.7%)	32 (44.4%)
76-85 years	37 (41.1%)	27 (37.5%)
N 85 years	11 (12.2%)	13 (18.1%)
Race - no. (%)			0.852
Chinese	75 (83.3%)	60 (83.3%)
Malay	4 (4.4%)	5 (6.9%)
Indian	7 (7.8%)	5 (6.9%)
Others	4 (4.4%)	2 (2.8%)
Medical history - no. (%)
NIL/unknown	4 (4.4%)	4 (5.6%)	0.532
Heart disease	39 (43.3%)	24 (33.3%)	0.256
Diabetes	32 (35.6%)	31 (43.1%)	0.337
Hypertension	66 (73.3%)	56 (77.8%)	0.584
Stroke	9 (10.0%)	5 (6.9%)	0.581
Glaucoma	2 (2.2%)	1 (1.4%)	0.696
Asthma	11 (12.2%)	4 (5.6%)	0.179
COLD/COPD	3 (3.3%)	1 (1.4%)	0.630
Psychiatric disorder	5 (5.6%)	1 (1.4%)	0.228
Renal disease	14 (15.6%)	7 (9.7%)	0.349
Osteoporosis	4 (4.4%)	4 (5.6%)	0.746
Epilepsy	0 (0%)	0 (0%)	-
Hyperlipidemia	54 (60.0%)	44 (61.1%)	0.886
Cancer	16 (16.7%)	6 (8.3%)	0.158
Others	40 (44.4%)	29 (40.3%)	0.634
Number of co-morbidities - no. (%)			0.610
0-1	14 (15.9%)	11 (15.7%)
2-3	30 (34.1%)	29 (41.4%)
N 3	44 (50.0%)	30 (42.9%)
Social history - no. (%)
Smoker	4 (4.4%)	6 (8.3%)	0.342
Heavy drinker (bingeing episodes or N 4 units/week)	0 (0%)	1 (1.4%)	0.444
Previously known to MSW	1 (1.1%)	1 (1.4%)	0.874
Known substance abuser	0 (0%)	1 (1.4%)	0.444
Unemployed	74 (82.2%)	55 (76.4%)	0.433
Lives alone	80 (88.9%)	55 (76.4%)	0.055
Bad debtor	5 (5.6%)	5 (6.9%)	0.752
Highest level of education - no. (%)			0.380
Primary and below	74 (82.2%)	53 (73.6%)
Secondary and post-secondary	12 (13.3%)	13 (18.1%)
Tertiary	4 (4.4%)	6 (8.3%)
marital status- no. (%)			0.199
Single	9 (10.0%)	6 (8.3%)
Married (or living as married)	32 (35.6%)	17 (23.6%)
Separated but not divorced/divorced/widowed	49 (54.4%)	49 (68.1%)
Type of housing - no. (%)			0.964
HDB 1- to 3-room	52 (57.8%)	41 (56.9%)
HDB 4- to 5-room/executive	31 (34.4%)	26 (36.1%)
Private flat/condominium/landed property	7 (7.8%)	5 (6.9%)
Total monthly income- no. (%)			0.345
No income	83 (92.2%)	70 (97.2%)
<=$3999	6 (6.7%)	2 (2.8%)
>=$4000	1 (1.1%)	0 (0%)
Chief complaint of initial visit - no. (%)			0.678
Pain	49 (54.4%)	37 (51.4%)
Vomiting	0 (0%)	1 (1.4%)
Diarrhoea	1 (1.1%)	1 (1.4%)
Seizures/fits	0 (0%)	0 (0%)
Weakness	2 (2.2%)	1 (1.4%)
Bleeding	1 (1.1%)	2 (2.8%)
Breathlessness	8 (8.9%)	8 (11.1%)
Altered mental state	0 (0%)	1 (1.4%)
Fever	0 (0%)	1 (1.4%)
Giddiness/dizziness	21 (23.3%)	11 (15.3%)
Others	8 (8.9%)	9 (12.5%)
Presenting complaint - no. (%)			0.117
Trauma	31 (34.4%)	16 (22.2%)
Non-trauma	59 (65.6%)	56 (77.8%)

Table 2

Comparing the number of Emergency Department visits and hospital admissions between Medical Alert Protection System arm and telephone follow-up arm, and between before and after intervention.

Outcome Telephone follow-up only

(N = 90)

Medical Alert Protection System (N = 72)

Telephone vs medical alert

p-value

	Pre-intervention	Post-intervention		Pre-intervention	Post-intervention
No. of ED visits; median, Interquartile Range(IQR)	1	0		1	0
	(1,2) [1,32]	(0,1) [0,36]		(1,2) [1,6]	(0,1) [0,14]
Difference in ED visits (pre - post); median (IQR)						0.881
[range] p-value	1			1
	(0,1) [-9.7]			(0,1) [-11,6]
	0.001			0.001
No. of hospital admissions; median (IQR)[range]	1	0		1	0
	(0,2) [0,15]	(0,1) [0,15]		(0,1) [0,4]	(0,1) [0,5]
Difference in hospital admissions (pre - post); median (IQR)[range] p-value	1 (0,1) [-4.6]			0 (0,1) [-5.3]		0.545
	0.002			0.002

were significant reductions in the median number of ED visits in the six month periods before, compared to after intervention (p = 0.001 for both). A breakdown of the cumulative Emergency Department Visits in the six months post-intervention between the study arms is shown in Fig. 2. There was no significant difference between the two arms (p = 0.523). A breakdown of the cumulative hospital admissions in the six months post-intervention between the study arms is shown in Fig. 3. There was no significant difference between the two arms (p = 0.104) (Table 3).

In terms of decrease in the median number of hospital admissions compared to the pre-intervention (baseline) phase, there was no signif- icant difference between the two study arms (p = 0.55). However, in both individual arms, there were significant reductions in the median number of hospital admissions in the six months period before com- pared to after intervention (p = 0.002 for both).

Among participants who have had one or more admissions during the six months period post intervention, the MAPS arm had significantly lower median total length of stay (8 days, Interquartile Range [IQR] = (4, 14)) compared to the control arm (15 days, IQR = (3, 25), p = 0.045).

Secondary outcomes

In terms of confidence scale scores (Table 4) at one week, one month and six months post intervention, there was no significant difference between the two arms for absolute scores at each measurement and there was no significant difference in the increase in confidence score from one week to six months.

In terms of EQ-5D self-reported 5 dimensions at six months post intervention (Table 4), the MAPS arm had significantly better outcomes in areas of “usual activities” (70.5% reporting “no problem” versus 44.9%,

Fig. 2. Comparison of cumulative Emergency Department Visits in the six months post- intervention between Medical Alert Protection System arm and telephone follow-up arm.

p = 0.007), “pain/discomfort” (72.4% reporting “no problem” versus 48.3%, p = 0.006), and “anxiety/depression” (82.3% reporting “no prob- lem” versus 61.8%, p = 0.014).

In addition, the median score for health state was significantly higher in the MAPS arm (70 IQR = (60,80) versus 60 IQR = (50,70), p = 0.008). There was no difference in mortality at six months post intervention between the two arms.

Discussion

In the population studied of home-alone elderly patients seen at the Emergency Department with at least low-moderate fall risk, the use of the MAPS resulted in a reduced total length of stay for patients with one or more admissions. There was no significant difference in terms of our first two primary outcome measures of Subsequent ED visits and hospital admissions between using MAPS and telephone follow- up only. Both arms experienced decreased ED visits over the period of follow up, with no difference between arms. This might indicate that telephone follow-up and having a number to call a nurse to clarify doubts might be as effective as a MAPS in reducing unwarranted ED visits. However it could also be a natural phenomenon of regression to the mean after acute episodes.

For those admitted during the prospective study period, we were able to show a reduced total length of stay for those on the device com- pared to those receiving only telephone follow-up. A possible explana- tion could be increased threshold for clinicians to discharge these elderly patients where previously they may be in doubt of the patients’

Fig. 3. Fig. 2. Comparison of cumulative hospital admissions in the six months post- intervention between Medical Alert Protection System arm and telephone follow-up arm.

Table 3 Differences in total length of stay for six months post intervention for patients with hospi- tal admissions, between Medical Alert Protection System arm and telephone follow-up arm.

daily activities due to fear of being unable to access help when they run into trouble [26].

While there was no significant difference in terms of subsequent ED

visits and hospital admissions between using MAPS and telephone fol-

Outcome Patients with hospital admissions from telephone follow-up arm

(N = 31)

Patients with hospital admissions Medical Alert Protection System arm (N = 23)

p-Value

low-up only, there were significant reductions of both variables when compared to a six-month pre-intervention baseline. This implies, even simple follow-up measures as phone calls at certain intervals post dis- charge might have allowed patients to access interim basic medical ad-

Total length of hospital stay per admitted patient; median

(IQR)[range]

15

(3,25) [1,77]

8

(4,14) [1,47]

0.045

vice early negating the need for ED visit for simple complaints, assuming there are no confounding developments that would result in this differ- ence over time. It is also possible that a larger sample size would power the differentiation of the two interventions in terms of these two outcomes.

This study has several limitations. First, due to practical differences in the interventions, there was no blinding of the participants. This

ability to seek assistance in situations of emergency. The increased con- fidence scale score found in this study especially in the domain of “usual activities” may have contributed to the patient component in joint deci- sions in discharge planning.

The higher quality of life scores (both EQ-5D as well as overall) in the MAPS arm was a significant finding. This signal affirms previous authors who postulated that the benefits of such devices are mediated by reduc- ing reluctance in elderly to independently ambulate or perform other

could lend to bias due to placebo effect for patients who may derive wellness simply from the novelty of the MAPS. However, this was de- signed as a real-world implementation trial, and it might be that this ef- fect is desirable in achieving certain subjective benefits such as increased quality of living even in actual implementation. Secondly, there might have been surveillance bias as patients in both arms may feel the need to be more judicious about their ED visits although they were not told specifically that their rate of ED use was being measured.

Table 4

Comparison of confidence scale scores, EQ-5D scores, and mortality rate between Medical Alert Protection System arm and telephone follow-up arm.

Secondary outcome			Telephone follow-up only (N = 90)		Medical Alert Protection System (N = 72)	p-Value
Confidence scale score, median (IQR)[range] (non-missing N)		At 1 week	5 (4, 6) [2, 8] (N = 81)		5 (5, 6) [2, 8] (N = 52)	0.310
		At 1 month At 6 months	6 (5, 7) [2, 8] (N = 87) 6 (5, 7) [3, 10]		6 (5, 7) [3, 9] (N = 64) 7(6, 8) [3, 9]	0.424 0.126
		Change from 1 week to 6 months, (6 m-1 w)	(N = 90) 1(0, 2) [-3, 4]		(N = 60) 2(0, 2.5) [-1, 6] (N = 44)	0.225
		p-value (6 m vs 1 w)	(N = 81) p = 0.001		p = 0.001
Secondary outcome		Telephone follow-up only (N = 89)		Medical alert protection system (N = 62)		p-Value
EQ-5D subscales at 6 months	Mobility No problem, n(%)	36 (40.4%)		31 (50.0%)		0.393
	Some problem, n(%)	47 (52.8%)		29 (26.8%)
	Problem, n(%) Self-care	6 (6.7%)		2 (3.2%)		0.116
	No problem, n(%)	47 (52.8%)		43 (69.4.0%)
	Some problem, n(%)	37 (41.6)		16 (25.8%)
	Problem, n(%) Usual activities No problem, n(%)	5 (5.6%) 40 (44.9%)		3 (4.8%) 43 (70.5%)		0.007
	Some problem, n(%)	37 (41.6%)		15 (24.6%)
	Problem, n(%) Pain/discomfort	12 (13.5%)		3 (4.9%)		0.006
	No problem, n(%)	43 (48.3%)		46 (72.4%)
	Some problem, n(%)	42 (47.2%)		14 (22.6%)
	Problem, n(%) Anxiety/depression No problem, n(%)	4 (4.5%) 51 (61.8%)		2 (3.2%) 51 (82.3%)		0.014
	Some problem, n(%)	32 (36.0%)		9 (14.5%)
	Problem, n(%)	2 (2.2%)		2 (3.2%)
	Health State, median (IQR)[range]	60 (50, 70) [10, 100]		70 (60, 80) [20, 100]		0.008
Secondary outcome		Telephone follow-up only (N = 90)	Medical Alert Protection System (N = 72)			p-Value
Mortality within 6 month post-intervention n(%)		0 (0%)	2 (2.8%)			0.112

Thirdly, the data lacked the granularity to discern whether the partici- pants used the MAPS appropriately.

Another major limitation is the relatively high dropout rate from the MAPS arm (34/106 patients, 32.0%). In particular, 22 patients dropped out due to having “changed their minds”. This is an issue that needs to be addressed in order to reduce wastage if such a system is to be imple- mented as a protocol in the ED. Indeed, adherence and uptake among the community has been found to be a problem in previous studies [16,27]. A survey of Australian MAPS users found use to be low especial- ly in the shower and in bed at night [16]. That study implicated as rea- sons the lack of knowledge of alarm functioning, distrust of technology and a device design that was uncomfortable or unaesthetic. There may have been some selection bias due to the dropouts and the subsequent need to adjust recruitment ratios. Also the study would have probably benefited from a longer follow up period, as 6 months may not be enough to show differences in subsequent ED visits and admissions. Unfortunately we could only support a 6 month follow-up, as manpower was limited to support this trial and we were only able to purchase part of the nurses’ time for the trial. Subsequent studies could focus on doing qualitative assessments of how the device

interacts with the patient in the home environment, and impacts the quality of life at home.

Conclusion

In the population studied of home-alone elderly patients seen at the Emergency Department with at least low-moderate fall risk, the use of the MAPS resulted in a reduced total length of stay for patients with one or more admissions. There was no significant difference in terms of subsequent ED visits and hospital admissions between using MAPS and telephone follow-up only. There was increased confidence scores and quality of life scores (EQ-5D and EQ-VAS) in the MAPS arm.

Disclosure statement

Funding was provided by SG Enable (previously Centre for Enabled Living), Ministry of Social and Family Development (HQIF 2012 / 01), Singapore for this study, as part of its Sustainable Enhancement for Eldercare and Disability Services Fund.

Appendix 1. Call log

Date & time: 1 week

The following questions are about your CURRENT situation:

How confident are you that you can get timely help during emergency at home?

Not at all										Completely
0	1	2	3	4	5	6	7	8	9	10

What do you LIKE about the E-Alert system? What do you DISLIKE about the E-Alert system?

Given a choice, would you like to continue using the E-Alert system?

Date & time: 1 month

The following questions are about your CURRENT situation:

How confident are you that you can get timely help during emergency at home?

Not at all										Completely
0	1	2	3	4	5	6	7	8	9	10

What do you LIKE about the E-Alert system? What do you DISLIKE about the E-Alert system?

Given a choice, would you like to continue using the E-Alert system?

Date & time: 6 months

The following questions are about your CURRENT situation:

How confident are you that you can get timely help during emergency at home?

Not at all										Completely
0	1	2	3	4	5	6	7	8	9	10

What do you LIKE about the E-Alert system? What do you DISLIKE about the E-Alert system?

Given a choice, would you like to continue using the E-Alert system?

Visits to the Hospital during 6 months of participation.

Date Diagnosis Disposition Length of stay

EQ5D at 6 months.

References

1. Department of Statistics Singapore. Population trends. https://www.singstat.gov.sg/ docs/default-source/default-document-library/publications/publications_and_ papers/population_and_population_structure/population2015.pdf; 2015. [Published 2015].
2. World Health Organization. World health statistics 2013 - global health indicators. http://www.who.int/gho/publications/world_health_statistics/EN_WHS2013_Full. pdf. [Published 2013].
3. Goh VHH. Aging in Asia: a cultural , socio-economical and historical perspective. Aging Male 2005;8(June):90-6. https://doi.org/10.1080/13685530500088472.
4. Singal BM, Hedges JR, Rousseau EW, Sanders AB, Berstein E, McNamara RM, et al. Ge- riatric patient emergency visits. Part I: comparison of visits by geriatric and younger patients. Ann Emerg Med 1992;21(7):802-7.
5. Lowenstein SR, Crescenzi CA, Kern DC, Steel K. Care of the elderly in the emergency department. Ann Emerg Med 1986;15(5):528-35.
6. Aminzadeh F, Dalziel WB. Older adults in the emergency department: a systematic review of patterns of use, adverse outcomes, and effectiveness of interventions. Ann Emerg Med 2002;39(3):238-47.
7. Foo CL, Siu VWY, Tan TL, Ding YY, Seow E. geriatric assessment and intervention in an emergency department observation unit reduced re-attendance and hospitalisation rates. Australas J Ageing 2012;31(1):40-6. https://doi.org/10.1111/ j.1741-6612.2010.00499.x.
8. Chik LF, Kim CC, Hsin KG, Seow E. Profiling acute presenting symptoms of geriatric patients attending an Urban Hospital Emergency Department. Ann Acad Med Singap 2009;38(6):515-20.
9. Zarisfi F, Hong Q, Seah P, Li H, Yap S, Ong M. Retrospective study of elderly Frequent attenders presenting with chest pain at emergency department. Int J Emerg Med 2014;7(1):35. https://doi.org/10.1186/s12245-014-0035-2.
10. Lim LL, Kua EH. Living alone, loneliness, and psychological well-being of older per- sons in Singapore. Curr Gerontol Geriatr Res 2011;2011. https://doi.org/10.1155/ 2011/673181.
11. Takahashi M, Aiba M, Sakurai T, Sakamoto N, Seok Yang K, Tsuda H, et al. A clinical analysis of elderly dementia patients with physical comorbidities. Nihon Ronen Igakkai Zasshi 2016;53(1):45-53. https://doi.org/10.3143/geriatrics.53.45.
12. Anantharaman V. Impact of health care system interventions on emergency depart- ment utilization and overcrowding in Singapore. Int J Emerg Med 2008;1(1):11-20. https://doi.org/10.1007/s12245-008-0004-8.
13. Koch WJ. Emergency response system assists in discharge planning. Dimens Health Serv 1984;61(11):30-1.
14. Roush RE, Teasdale TA, Murphy JN, Kirk MS. Impact of a personal emergency re- sponse system on hospital utilization by community-residing elders. South Med J 1995;88(9):917-22.
15. Lee JS, Hurley MJ, Carew D, Fisher R, Kiss A, Drummond N. A randomized clinical trial to assess the impact on an emergency response system on anxiety and Health care use among older emergency patients after a fall. Acad Emerg Med 2007;14(4): 301-8. https://doi.org/10.1197/j.aem.2006.11.017.
16. Miguel KDS, Lewin G. Personal emergency alarms: what impact do they have on older people‘s lives? Australas J Ageing 2008;27(2):103-5. https://doi.org/10. 1111/j.1741-6612.2008.00286.x.
17. Statistics Singapore - Department of Statistics Singapore. http://www.singstat.gov. sg/statistics/latest_data.html#14.
18. Ho AFW, Chew D, Wong TH, Ng YY, Pek PP, Lim SH, et al. prehospital trauma care in Singapore. Prehosp Emerg Care 2014:1-7. https://doi.org/10.3109/10903127.2014. 980477 [Epub ahead].
19. Herdman M, Gudex C, Lloyd A, Janssen M, Kind P, Parkin D, et al. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res 2011;20(10):1727-36. https://doi.org/10.1007/s11136-011-9903-x.
20. Luo N, Wang P, Thumboo J, Lim YW, Vrijhoef HJM. Valuation of EQ-5D-3L health states in Singapore: modeling of time trade-off values for 80 empirically observed health states. PharmacoEconomics 2014;32(5):495-507. https://doi.org/10.1007/ s40273-014-0142-1.
21. Wang Y, Tan N-C, Tay E-G, Thumboo J, Luo N. Cross-cultural measurement equiva- lence of the 5-level EQ-5D (EQ-5D-5L) in patients with type 2 diabetes mellitus in Singapore. Health Qual Life Outcomes 2015;13:103. https://doi.org/10.1186/ s12955-015-0297-2.
22. Luo N, Wang Y, How CH, Tay EG, Thumboo J, Herdman M. Interpretation and use of the 5-level EQ-5D response labels varied with survey language among Asians in Sin- gapore. J Clin Epidemiol 2015;68(10):1195-204. https://doi.org/10.1016/j.jclinepi. 2015.04.011.
23. Luo N, Chew LH, Fong KY, Koh DR, Ng SC, Yoon KH, et al. Validity and reliability of the EQ-5D self-report questionnaire in Chinese-speaking patients with rheumatic diseases in Singapore. Ann Acad Med Singap 2003;32(5):685-90.
24. O’Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics 1979;35(3):549-56.
25. Hintze JPASS. NCSS, LLC Kaysville, Utah, USA; 2008.
26. Vellas B, Cayla F, Bocquet H, de Pemille F, Albarede JL. Prospective study of restriction of activity in old people after falls. Age Ageing 1987;16(3):189-93.
27. NYMAN SR, VICTOR CR. Use of personal call alarms among community-dwelling older people. Ageing Soc 2012;34(01):1-23. https://doi.org/10.1017/ S0144686X12000803.