Rural-Urban Comparison of Telehome Monitoring for Patients with Chronic Heart Failure

Abstract

Introduction: Rural geographic isolation may act as a promoting or restraining variable to the diffusion of technology and healthy aging in the community. Telehome monitoring (TM) leverages technology to support seniors living in the community with chronic conditions. To date, limited research has investigated the utilization of TM in rural settings. This study assesses the comparative utilization of TM for patients with heart failure in rural versus urban environments.

Materials and Methods: We conducted a cross-sectional study involving chart reviews of all patients enrolled in the TM program at the University of Ottawa Heart Institute during 2014. Data were extracted on urban/rural status, demographic characteristics, and process and outcomes of care. Descriptive, bivariate, and multivariate analyses were conducted.

Results: More rural patients did not have a documented reason for emergency room visits compared to urban patients. There was no significant association between the urban/rural status and the process and outcome measures at the multivariate level. Being followed-up regularly by a family physician and a specialist, as opposed to a specialist or general practitioner only, was associated with significantly longer TM period and a higher number of diuretic adjustments and calls made by nurses.

Discussion: Although more urban patients were older and living alone, their profile did not affect their utilization of TM. The difference in diagnosis between urban and rural patients also did not contribute to such differences. Hence, there is no variation in the process and outcome measures associated with the utilization of TM between urban and rural environments.

Conclusions: Rural patients may not be perceived as extensive users of resources nor patients who represent challenges in terms of feasibility of TM use.

Introduction

e-Health is a term that is often used in today's healthcare environment to refer to the application of technologies and the leveraging of the internet for the delivery of information and services in the health sector.^1,2 With the worldwide phenomenon of population aging, and the dearth of resources available to address the challenges associated with an increasing prevalence of chronic conditions, e-health tools are becoming increasingly popular. Specifically, e-health consumer applications are considered as important tools for supporting self-management and self-care for patients with chronic conditions in the community.³

Telehome monitoring (TM) represents one of the consumer e-health applications that have gained increased attention over the years. It refers to the remote and noninvasive monitoring of patients based on automated transmission of physiological and clinical data related to their conditions, from their place of residence to a healthcare organization, which provides ongoing follow-up and treatment.⁴ TM has been used to support the management of patients with a broad range of chronic diseases, and its promising impacts have been documented in relation to various chronic health conditions, including heart failure (HF), as confirmed by numerous systematic reviews published in this area of research (e.g., Paré et al.,⁴ Jaana and Paré,⁵ Jaana et al.,^6,7 and Kitsiou et al.⁸). A systematic review by Pare et al.⁴ on TM and chronic diseases (including HF, pulmonary conditions, diabetes, and hypertension) reported general patients' compliance and satisfaction with TM programs and with the overall use of this technology.⁴ The authors also indicated improvement in clinical effectiveness outcomes (e.g., including decrease in emergency visits, hospital admissions, and average length of stay), mostly for heart and diabetes conditions. In a more recent overview of systematic reviews on TM interventions and HF, specifically, Kitsiou et al.⁸ further demonstrated the positive impacts of TM on reducing the relative risk of all-cause mortality and HF-related hospitalizations, compared to usual care. The systematic reviews that they considered in this critical overview of the literature consistently documented positive effects of TM on these outcome measures, as well as satisfaction of patients with this approach/technology.⁸ Despite this evidence, limited information was available on the effectiveness of TM across different groups and types of users. Kitsiou et al.⁸ recommended that future studies identify optimal strategies for TM success and examine whether the effectiveness of TM interventions varies between diverse patient groups, as brought up by the reviews they considered.

Recent studies on e-health have focused on the factors that affect the use of various applications, in general, among patients (e.g., de Veer et al.,³ Or et al.,⁹ and Cimperman et al.¹⁰), which may lead to differential impacts and benefits for them. Prior research on change implementation has stressed on the importance of considering the “context” when explaining the outcomes related to a change.^11,12 Specifically, when investigating the outcomes of an e-health application, it is necessary to consider the environment/context in which the implementation takes place. This context varies based on the geographical location of the patients (e.g., urban versus rural), as manifested by a variation in the social environment, individual resources and capabilities, and need for the technology.¹¹ Along the same line, a recent systematic review by Peek et al.¹³ on the factors that influence the acceptance of technology for aging in place identified factors related to six themes that affect technology acceptance: concerns about technology; expected benefits; technology need; social influence; and characteristics of patients. Hence, we hypothesize in this study that these factors will differ between patients residing in urban versus rural environments, and subsequently, the use and impacts of TM will vary between these two groups.

Rural e-health is an emerging area of research, which has not fully matured and developed.¹¹ In a systematic literature review on e-health adoption, Hage et al.¹¹ found that geographic isolation in rural settings (i.e., geographical areas with low population density, limited resources, and relative isolation) may act as both a promoting, as well as a restraining, variable for e-health. They discussed this context as being characterized by existing individual and socioeconomic structures (e.g., aging and demographics, education, familiarity with technologies, and isolated areas), which may slow the diffusion of e-health in rural communities.¹¹ Yet, the rural context also presents an opportunity for e-health to make an impact and replace services that are missing in small communities or complement existing ones.¹¹ Another review by Warburton et al.¹⁴ on the adoption of information and communication technologies (ICTs) in rural communities in Australia discussed local and international evidence that indicate the potential of technology to promote healthy aging in disadvantaged rural communities. In addition, Jones and Grech¹⁵ synthesized the evidence on HF patients' experience with TM in rural settings, in specific, and revealed gaps in the literature in terms of studies, including patients with this condition who live in rural and remote locations and may benefit from TM. Although a few studies have previously explored telehealth applications in rural environments in relation to patients' education,¹⁶ technology use by nurses during home visits,¹⁷ attitude toward telehealth devices,¹⁸ feasibility for cancer patients,¹⁹ and interactive telephone support,²⁰ limited research exists on TM impacts in rural versus urban settings. To our knowledge, only one study has specifically assessed the utilization outcomes of TM in rural settings in the United States and reported decreased odds of acute hospitalizations and increased odds of discharge to the community among patients with multiple chronic conditions.²¹

This study addresses this gap and assesses the comparative utilization of this patient management approach for patients living in rural versus urban environments. Specifically, we focus on the process of care (e.g., adjustment of medications and adjustment of diuretics) and outcomes related to the utilization of services (e.g., emergency visits and admissions) among these two groups of patients.

Materials and Methods

The University of Ottawa Heart Institute (UOHI) has a well-established TM program. Any inpatient with a discharge diagnosis of congestive heart failure is automatically referred for TM. An advanced practice nurse reviews the referral against standardized acceptance criteria, and the patient is either placed in the TM or automated calling program based on their clinical needs and ability to use the technology. This cross-sectional study examines the differences in the utilization of TM among patients living in urban and rural settings. For this purpose, we conducted a chart review of all patients enrolled in the TM program at the UOHI, which serves the capital of Ottawa and the Champlain region, during the year 2014. The TM system used at the UOHI consists of a user-friendly monitor connected to a cellular or telephone line. It provides daily data transmission from standardized peripherals (direct download), which are part of the monitoring system, including weigh scales, blood pressure cuffs, and ECGs (electrocardiography). The data transmitted by the patients are received at a central TM station at the UOHI and reviewed by a nurse expert who intervenes based on standardized protocol (e.g., medication titration to optimal doses) and communicates with the patients as needed to avoid deterioration in their conditions. They may also be in contact with family physicians or specialists as required.

In addition to the importance of population size in determining the rural status, there are other attributes (e.g., distance to an essential service, population density, and so on) that are also relevant to consider. According to Statistics Canada,²² the definition of “rural” status should be determined based on the question that we are addressing. In their systematic review on e-health adoption in rural communities, Hage et al.¹¹ described the rural context as a geographical area that has a low population density, limited resource bases, and relative isolation and social homogeneity. This definition, which extends beyond the geographic concept to include a social dimension, is relevant in this research related to e-health in the community. Hence, and given the context of this study, we used a classification of urban/rural status based on categories proposed by Statistics Canada.²³ Specifically, we defined “rural” status as the home residence of the patient being in a geographic area/community consisting of remote rural areas of <1,000 habitants, as well as small population centers of 1,000–29,999 persons. The latter often lack the resources available in medium and larger population centers and have more limited access to specialized care. In this specific research, these represent referring areas for specialized HF care delivered at the UOHI. “Urban” status therefore refers to the home residence of the patient being in a geographic area/community with a population ≥30,000 persons.

The measures related to the process of care included: (1) frequency of times the vital signs were out of range for a patient (e.g., high/low blood pressure and high/low pulse); (2) frequency of calls made to a patient by the nurse; (3) frequency of changes to diuretic doses in response to change in weight and HF symptoms; and (4) frequency of changes to HF medications based on best practice (e.g., Angiotensin Converting Enzyme Inhibitor [ACEI] and Mineralocorticoid Receptor Antagonist [MRA]). The measures related to the outcome of care included the following: (1) frequency of emergency room (ER) visits by a patient; (2) frequency of hospital admissions by a patient; and (3) incidence of death.

Descriptive analyses provided an overview of the sample characteristics. Bivariate analyses were conducted to assess the significant association among the urban/rural status, patients' characteristics, and the process and outcome measures considered in this study. Specifically, chi square was used to assess the significant associations between categorical variables at the bivariate level. Correlation analysis was used to test the significant association between continuous variables (e.g., age and process of care measures), and t test and analysis of variance were used to assess the significant relationships between independent variables with two or more categories, respectively, and the continuous process and outcomes of care measures. Finally, multivariate regression analyses included the variables that showed significant relationships at the bivariate level.

Results

Sample Characteristics

A total of 240 patients (43% rural and 57% urban) were included in the TM program in the year 2014. The mean age of the sample was 72 years (range = 30–96 years); patients living in urban areas were a bit older than their counterparts in rural areas. On average, patients were followed for a duration of 108 days (range = 3–427 days). As shown in Table 1, more than half were men and did not live alone; the rural group had a significantly higher proportion of patients not living alone compared to the urban group. A large number of patients (66%) were diagnosed with systolic heart failure (SHF), which is a serious condition that may vary in degree of severity, and had a high to very high level of comorbidities (63%) based on the Charlson Comorbidity Index (CCI).²⁴ The rural group had a significantly higher proportion of patients diagnosed with diastolic heart failure (DHF), which is usually observed among elderly and diabetic patients and is associated with long-term hypertension. The majority of patients (82%) reported both a general practitioner (GP) and a specialist as regular providers; 12% and 6% were only seen by either a GP or a specialist only, respectively.

Table 1.

General Patients' Characteristics

	URBAN	RURAL	TOTAL
	N (%)	N (%)	N (%)
Gender
Male	86 (63)	69 (67)	155 (65)
Female	51 (37)	34 (33)	85 (35)
Living condition^a
Alone	55 (40)	30 (29)	85 (35)
Not alone	82 (60)	73 (71)	155 (65)
Diagnosis^a
Systolic heart failure	96 (70)	62 (60)	158 (66)
Diastolic heart failure	34 (25)	39 (38)	73 (30)
Unspecified	7 (5)	2 (2)	9 (4)
Comorbidities (CCI)^b
Low	12 (9)	16 (15)	28 (12)
Moderate	35 (25)	27 (26)	62 (26)
High	53 (39)	42 (41)	95 (40)
Very high	37 (27)	18 (17)	55 (23)
Regular provider^c
GP only	12 (9)	17 (16)	29 (12)
Specialist only	8 (6)	7 (8)	15 (6)
Both GP and specialist	117 (85)	79 (77)	196 (82)

	MEAN (MINIMUM TO MAXIMUM)	MEAN (MINIMUM TO MAXIMUM)	MEAN (MINIMUM TO MAXIMUM)
Age (years)^d	74 (30–96)	71 (30–96)	72 (30–96)
Duration of telemonitoring (days)	105 (7–300)	113 (3–427)	108 (3–427)

Significant differences between the urban and rural groups at 10% significance level.

Age-modified CCI: low: ≤3; moderate: 4–5; high: 6–7; very high: ≥8.

Patients under the care of a GP or specialist (i.e., a cardiologist or an internist) or both.

Significant differences between the urban and rural groups at 5% significance level.

CCI, Charlson Comorbidity Index; GP, general practitioner.

Variation in Process and Outcome Measures by Patients' Characteristics

Table 2 presents the results of the bivariate analyses examining the significant relationships between patients' characteristics and the process and outcome measures considered in this study. Gender and living condition (i.e., alone or not) were significantly associated with the number of ER visits at 10% significance level. Specifically, the mean number of ER visits for patients living alone was 0.18 (range = 0–3) compared to 0.30 (range = 0–3) for patients not living alone. Similarly, the mean number of ER visits for female patients was 0.15 (range = 0–2) compared to 0.31 for male patients (range = 0–3). The type of diagnosis was also significantly associated with the event of death, with a higher proportion of patients dying who have a DHF diagnosis (53%) as opposed to a SHF diagnosis (40%).

Table 2.

Relationships (p-Values) Between Patients' Characteristics and Process and Outcomes of Care Measures

PROCESS AND OUTCOMES	AGE^a	GENDER^b	LIVING CONDITION^b	DIAGNOSIS^b	COMORBIDITIES^b	PROVIDER^b
Duration of TM	0.509	0.364	0.921	0.885	0.055	0.008
No. of adjustments of diuretics and provider	0.658	0.527	0.917	0.513	0.994	0.068
No. of adjustments of HF best practice medications	0.625	0.339	0.813	0.244	0.932	0.390
No. of times vital signs were out of range	0.928	0.739	0.171	0.661	0.457	0.149
No. of patient calls by RN	0.634	0.442	0.724	0.232	0.136	0.008
No. of ER visits	0.834	0.040	0.086	0.353	0.774	0.067
No. of hospital admissions	0.401	0.906	0.956	0.699	0.961	0.511
Death^c	0.787	0.464	0.197	0.050	0.397	0.881

The p-values presented are associated with Pearson correlation coefficient between age and the process and outcomes, except for death (t test).

The p-values presented are associated with t test for gender and living condition; ANOVA test for diagnosis, comorbidities, and provider.

The p-values presented for the reason for ER visits, death, and cause of death in relation to all patient characteristics are associated with the Chi-square test.

p-Values < 0.10 are highlighted in bold indicating significant associations at 10% significance level.

ANOVA, analysis of variance; ER, emergency room; HF, heart failure; RN, registered nurse; TM, telehome monitoring.

The type of providers (GP and specialist, GP alone, and specialist alone) was significantly associated with the duration of TM, the number of adjustments in diuretics, the number of patient calls made by the nurse, as well as the number of ER visits. Patients who had both a GP and a specialist had the longest TM duration in days (mean = 113.95; range = 7–427), the highest number of diuretic adjustments (mean = 3.03; range = 0–51), the largest number of calls made by nurses (mean = 11.17; range = 1–36), and the highest number of ER visits (mean = 0.29; range = 0–3). Patients followed by only a GP came second with a mean TM duration of 87.72 days (range = 3–195), average number of diuretic adjustments of 1.48 (range = 0–11), and a mean number of calls made by nurses of 7.72 (range = 1–32). Interestingly, the average number of ER visits for these patients was the lowest among the three groups (mean = 0.03; range = 0–1). Patients who only had a specialist as a regular provider were followed for an average of 77.13 days (range = 15–168 days) and had the lowest number of diuretic adjustment (mean = 1.07; range = 0–4) and patient calls (mean = 7.60; range = 2–15); yet, they showed the second highest average ER visits of 0.20 (range = 0–1).

Relationship Between Urban/Rural Status of Patients and Process and Outcome Measures

A preliminary assessment of the relationship between the urban/rural status of patients and the process and outcome measures considered in this study ( Table 3 ) only revealed significance with the reason for ER visits (although data on this variable was missing for the majority of the patients). A larger number of patients in the rural group did not have on record a documented reason for ER visits compared to the urban group that had similar number of patients reporting to the ER for cardiac, noncardiac, and HF reason.

Table 3.

Bivariate Analysis of Urban/Rural Status in Relation to Process and Outcomes of Care Measures

PROCESS AND OUTCOMES	URBAN, N = 137	RURAL, N = 103	t TEST p-VALUE
	MEAN (MINIMUM TO MAXIMUM)	MEAN (MINIMUM TO MAXIMUM)
Process measures
No. of adjustments of diuretics	2.99 (0–51)	2.37 (0–15)	0.285
No. of adjustments of HF best practice medications	1.22 (0–27)	0.84 (0–11)	0.231
No. of times vital signs were out of range	0.75 (0–8)	0.78 (0–5)	0.874
No. of patient calls by RN	10.82 (1–36)	10.16 (1–35)	0.452
Outcome measures
No. of ER visits	0.30 (0–3)	0.19 (0–2)	0.152
No. of hospital admissions	0.39 (0–2)	0.38 (0–3)	0.921

	N (%)	N (%)	CHI SQUARE TEST p -VALUE
Death
Yes	8 (6)	7 (7)	0.762
No	129 (94)	96 (93)
Cause of death
Cardiac	3 (2)	1 (1)
Noncardiac	—	—	0.798
HF	1 (1)	1 (1)
Unknown	4 (3)	5 (5)
Not applicable	129 (94)	96 (93)

The multivariate analysis that we conducted, after controlling for the variables that showed significance at the bivariate level, did not reveal significant association between the urban/rural status of patients and the process and outcome measures considered in this study. Interestingly, the only significant relationship that persisted between these measures was the type of provider that patients had. Specifically, being followed-up regularly by a family physician and a specialist, as opposed to a specialist or GP only, was associated with significantly longer TM period (β = 28.4, p = 0.06), more ER visits (β = 0.2, p = 0.04), and a higher number of diuretic adjustments (β = 1.5, p = 0.09) and calls made by nurses (β = 3.4, p = 0.01). A post hoc analysis examining the correlation between the process and outcome measures, considered in this analysis ( Table 4 ), revealed significant correlations between the process measures of care (e.g., number of diuretic and medication adjustments and calls made by nurses) and the outcome measures of ER visits and hospital admissions. The only two exceptions were the frequency of changes in HF medications and the frequency of times the patient vital signs were out of range. These two measures were not positively correlated with the number of hospital admissions, which is a demonstration of the positive impact of TM in terms of early intervention and reducing hospital admissions.

Table 4.

Post Hoc Analysis of the Correlation Between Process and Outcome Measures

PROCESS AND OUTCOMES	DURATION	DIURETICS	MEDS	VS	CALLS	NO. OF ER	NO. OF ADM
Duration of TM (Duration)	1.00
No. of adjustments of diuretics and provider (Diuretics)	0.214^a	1.00
No. of adjustments of HF best practice medications (Meds)	0.133^b	0.639^a	1.00
No. of times vital signs were out of range (VS)	0.088	0.127^b	0.099	1.00
No. of patient calls by RN (calls)	0.475^a	0.510^a	0.278^a	0.321^a	1.00
No. of emergency room visits (No. of ER)	0.062	0.393^a	0.341^a	0.233^a	0.271^a	1.00
No. of hospital admissions (No. of Adm)	0.104	0.223^a	0.082	0.115	0.289^a	0.311^a	1.00

Significant at 0.01.

Significant at 0.05.

Discussion

Recent reviews have examined the potential of ICTs in promoting healthy aging within rural disadvantaged communities and discussed the effect of geographical isolation and social structures and inequalities in relation to the diffusion of e-health^11,14. Nevertheless, limited studies have included patients with HF in rural settings and assessed TM feasibility and usage among this group.¹⁵ This is especially relevant in light of the expected differences in patients' characteristics, concerns about technology, its need and benefits, and social influence compared to urban settings. This study contributes to this under-researched area and presents a comparison in TM utilization pattern among patients with HF living in urban versus rural settings.

The results revealed no variation in the process and outcome measures associated with the utilization of TM as a patient management approach between urban and rural settings. Diuretic adjustments, cardiac medication changes, and the number of calls made by nurses did not vary between urban and rural patients. The TM duration was also similar between the two groups. Although more patients were older and living alone in urban areas, their profile did not affect the pattern of utilization of TM nor the process and outcomes associated with it. Similarly, the difference in diagnosis, with more rural patients diagnosed with DHF as opposed to SHF, did not contribute to any variation between the two groups in relation to the process of care. As such, patients in rural areas do not require more resources or additional interventions when using TM.

The major variation in TM period, number of emergency visits, diuretic adjustments, and calls made by nurses was associated with the type of provider(s) regularly following patients, which persisted after controlling for all relevant variables. Surprisingly, patients who reported having a regular family physician and a specialist appear to be the highest consumer of resources, followed by patients who reported only having a family physician although the latter had the lowest emergency visits. This study was not designed, nor powered, to address this issue or deal with variations in the use of resources. Future studies should further investigate this issue to identify the reasons for this variation, which is especially relevant in the context of health systems as in Canada where family physicians play a major role as gatekeepers. In addition, future research may benefit from a longitudinal design to investigate the implications of TM on HF patients in the community in relation to the improvement in their ability for self-care management and quality of living.

It is important to acknowledge the limitations of this study, mostly in relation to the sample size and secondary nature of the data. Given the fact that quantitative data were already collected, we were limited in the analysis to the available variables and were not able to consider qualitative data; these may have provided complementary information useful to better understand the context. We were also unable to conduct any risk adjustment to the outcome measures given the limited information on the patients case-mix and complexity. In addition, the sample size was limited to the number of patients who were monitored at the one site. Despite these limitations, our analysis shows that TM is invaluable for patients living in rural areas where resources are scarce, and timely interventions may be more challenging. It improves their access to specialist HF care without needing to travel long distances and presents an opportunity to care for these patients while staying in the community.

In Ontario, there are challenges associated with people not living in urban areas, which limit their access to resources and services. In a document developed by the Ministry of Health and Long Term Care,²⁵ an action plan was proposed to change and improve the health system in the province. Among its pillars were providing faster access to appropriate care, including specialist care, and delivering integrated care in the community. The report also recommended leveraging technology to support the provision of these services.²⁵ The findings of this study inform the planning and development of services in this area. Since no significant differences were observed in relation to urban/rural status, rural patients may not be perceived as extensive users of resources nor patients who present challenges in terms of feasibility of TM use. Province-wide initiatives in support of this plan can leverage TM technology to provide specialized HF services to a growing population of elderly in the community, in rural and urban areas alike. In addition, providing TM services to patients living outside major city boundaries is a feasible approach to improve access to timely specialist services, which has been demonstrated to be associated with reduced HF-related hospitalizations.⁸ This is especially relevant in the context of the Canadian healthcare system that faces persisting challenges associated with emergency long wait time and limited access to healthcare services.²⁶

Footnotes

Acknowledgments

The authors thank Christine Struthers (APN Chronic Cardiac Care) for her support and assistance in planning and executing this project.

Disclosure Statement

No competing financial interests exist.

References

Eysenbach

What is e-health?. J Med Internet Res, 2001; 3:e20.

Health Canada. eHealth—Health Care System—Health Canada. Health Canada. 2010. Available at www.hc-sc.gc.ca/hcs-sss/ehealth-esante/index-eng.php (last accessed January 23, 2018 )

de Veer

, Peeters

, Brabers

, Schellevis

, Rademakers

JJJ

, Francke

. Determinants of the intention to use e-Health by community dwelling older people. BMC Health Serv Res, 2015; 15:103.

Paré

, Jaana

, Sicotte

. Systematic review of home telemonitoring for chronic diseases: The evidence base. J Am Med Inform Assoc, 2007; 14:269–277.

Jaana

, Paré

. Home telemonitoring of patients with diabetes: A systematic assessment of observed effects. J Eval Clin Pract, 2007; 13:242–253.

Jaana

, Paré

, Sicotte

. Hypertension home telemonitoring: Current evidence and recommendations for future studies. Dis Manag Health Outcomes, 2007; 15:19–31.

Jaana

, Paré

, Sicotte

. Home telemonitoring for respiratory conditions: A systematic review. Am J Manag Care, 2009; 15:313–320.

Kitsiou

, Paré

, Jaana

. Effects of home telemonitoring interventions on patients with chronic heart failure: An overview of systematic reviews. J Med Internet Res, 2015; 17:e63.

CKL

, Karsh

B-T

, Severtson

, Burke

, Brown

, Brennan

. Factors affecting home care patients' acceptance of a web-based interactive self-management technology. J Am Med Inform Assoc, 2011; 18:51–59.

10.

Cimperman

, Makovec Brenčič

, Trkman

. Analyzing older users' home telehealth services acceptance behavior—Applying an extended UTAUT model. Int J Med Inf, 2016; 90:22–31.

11.

Hage

, Roo

, van Offenbeek

, Boonstra

. Implementation factors and their effect on e-Health service adoption in rural communities: A systematic literature review. BMC Health Serv Res, 2013; 13:19.

12.

Pettigrew

, Whipp

. Managing change for competitive success. Oxford, UK and Cambridge, MA: B. Blackwell, 1991.

13.

Peek

STM

, Wouters

EJM

, van Hoof

, Luijkx

, Boeije

, Vrijhoef

HJM

. Factors influencing acceptance of technology for aging in place: A systematic review. Int J Med Inf, 2014; 83:235–248.

14.

Warburton

, Cowan

, Bathgate

. Building social capital among rural, older Australians through information and communication technologies: A review article: ICTs building social capital of rural older people. Australas J Ageing, 2013; 32:8–14.

15.

Jones

, Grech

. The patient experience of remote telemonitoring for heart failure in the rural setting: A literature review. Contemp Nurse, 2016; 52:230–243.

16.

Schweickert

, Rutledge

, Cattell-Gordon

, et al. Telehealth stroke education for rural elderly Virginians. Telemed J E Health, 2011; 17:784–788.

17.

Terschüren

, Fendrich

, van den Berg

, Hoffmann

. Implementing telemonitoring in the daily routine of a GP practice in a rural setting in northern Germany. J Telemed Telecare, 2007; 13:197–201.

18.

Terschüren

, Mensing

, Mekel

. Is telemonitoring an option against shortage of physicians in rural regions? Attitude towards telemedical devices in the North Rhine-Westphalian health survey, Germany. BMC Health Serv Res, 2012; 12:95.

19.

Petitte

, Narsavage

, Chen

Y-J

, Coole

, Forth

, Frick

. Feasibility study: Home telemonitoring for patients with lung cancer in a mountainous rural area. Oncol Nurs Forum, 2014; 41:153–161.

20.

Krum

, Forbes

, Yallop

, et al. Telephone support to rural and remote patients with heart failure: The chronic heart failure assessment by telephone (CHAT) study. Cardiovasc Ther, 2013; 31:230–237.

21.

Golbeck

, Hansen

, Lee

, Noblitt

, Christner

, Pinsonneault

. Telemonitoring improves home health utilization outcomes in rural settings. J Telemed Telecare, 2011; 17:273–278.

22.

Du Plessis

, Beshiri

, Bollman

, Clemenson

Definitions of rural. Ottawa: Statistics Canada, 2001. Available at www.statcan.gc.ca/pub/21-006-x/21-006-x2001003-eng.pdf (last accessed January 23, 2018 ).

23.

Statistics Canada. Population centre and rural area classification. 2016. Statistics Canada, 2016. Available at www.statcan.gc.ca/eng/subjects/standard/pcrac/2016/introduction (last accessed January 23, 2018 ).

24.

Charlson

, Pompei

, Ales

, MacKenzie

. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis, 1987; 40:373–383.

25.

Government of Ontario. Patients first: Action plan for health care. Ontario Ministry of Health and Long-Term Care, 2015. Available at www.health.gov.on.ca/en/ms/ecfa/healthy_change/docs/rep_patientsfirst.pdf (last accessed January 23, 2018 ).

26.

CIHI. Emergency room waits continue to increase. Canadian Institute for Health, 2017. Information. Available at www.cihi.ca/en/emergency-room-waits-continue-to-increase (last accessed January 23, 2018 ).