Effectiveness of IT-based interventions on self-management in adult kidney transplant recipients: a systematic review

Background Kidney transplant outcomes are broadly associated with transplant recipients’ capacity in following a complex and continuous self-management regimen. Health information technology has the potential to empower patients. This systematic review aimed to determine the impacts of IT-based interventions for self-management in kidney transplant recipients. Methods A comprehensive investigation was performed in MEDLINE (via PubMed) and EMBASE (via Scopus) in April 2019. Eligible studies were the randomized controlled trials which aimed to design an automated IT-based intervention. All English papers including adult kidney transplant recipients were included. To assess the clinical trial’s quality, Cochrane Collaboration’s assessment tool was employed. The articles were integrated based on category of outcomes, characteristics of interventions, and their impact. The interventions were classified based on the used IT-based tools, including smart phones, coverage tools, computer systems, and a combination of several tools. The impact of interventions was defined as: (1) positive effect (i.e. statistically significant), and (2) no effect (i.e. not statistically significant). Results A total of 2392 articles were retrieved and eight publications were included for full-text analysis. Interventions include those involving the use of computerized systems (3 studies), smart phone application (3 studies), and multiple components (2 studies). The studies evaluated 30 outcomes in total, including 24 care process and 6 clinical outcomes. In 18 (80%) out of 30 outcomes, interventions had a statistically significant positive effect, 66% in process and 33% in clinical outcomes. Conclusions IT-based interventions (e.g. mobile health applications, wearable devices, and computer systems) can improve self-management in kidney transplant recipients (including clinical and care process outcomes). However, further evaluation studies are required to quantify the impact of IT-based self-management interventions on short- and long-term clinical outcomes as well as health care costs and patients' quality of life.

their quality of life, and reduces the risk of mortality in the final stage of the disease [1][2][3][4][5]. In case of a transplant rejection, mortality rates and health care costs increase [6]. Given the limited health care resources due to growing demand, it is essential to create effective, non-interventional methods to promote self-management and monitor transplant recipients in order to increase the chances of success [7].
Transplant outcomes are broadly associated with transplant recipients' capacity in following a complex and continuous self-management regimen, thus minimizing the risks of transplant rejection and related diseases [8,9]. The ability to manage the outcomes of a chronic disease is defined as self-care. Self-management is classified into three groups: focusing on the disease needs; activating resources and living with chronic disease; and finally, managing drugs, roles, and feelings [10]. For kidney transplant recipients, self-management tasks include accepting medications, monitoring signs and symptoms of transplant rejection, performing regular periodic visits by specialists, compatibility with changes in social roles and communications, managing feelings, and creating new perspectives in life [11][12][13]. There is a need for more effective strategies to empower patients for self-management [14,15].
The advent of health information technology (IT) and its related instruments have shown potential advantages for patients to actively participate in their health monitoring as well as assistance for health care providers [16]. Information technology has created new ways for providing health care and training to patients. Such improvements provide a basis for redesigning health care processes based on the use and integrity of electronic communications at all levels. It has been previously shown that information technology can empower patients. This technology is transforming the patient's role from an inactive care services recipient into an active role in which patients are aware and involved in decisionmaking processes and have the right to choose [17]. In fact, a large number of researchers and system designers who previously focused on designing information technology software for care providers have now moved towards designing patient-focused software solutions [18].
Majority of interventional studies have examined the use of IT in designing patient training and monitoring for self-management in chronic diseases, including kidney transplant recipients. Some of these studies were effective, while others were ineffective. These heterogeneous results make it necessary to conduct a systematic review aiming to abstract the results of published studies. Several systematic reviews have been published on self-management patients with chronic kidney disease [19][20][21] and there is even a systematic study conducted on ITbased interventions to measure the self-management indices among patients with chronic renal disorders [22].
Although IT-based interventions have the potential to inform patients and improve self-care, there is a need for a provider to intervene in non-automated IT-based interventions; however, this would entail a potential source of human error. IT-based interventions can reduce error, improve self-management, and increase patients' awareness and knowledge. None of the automated ITbased systematic reviews had focused on kidney transplant recipient patients while patients' self-management behaviors will help them to improve clinical and process outcomes both before and after transplantation. In fact, this need begins before the transplantation, and as long as they continue to live with the transplanted kidney, this basic need continues to be present. Therefore, this study was conducted to describe the main features of the ITbased interventions and summarize their effectiveness on self-management outcomes among kidney transplant recipients. The main questions to be addressed on this topic include: What IT-based interventions studies have been conducted on kidney transplant recipients? What are the main characteristics of these studies? Did these interventions result in positive improvement on selfmanagement outcomes (clinical and process care) among kidney transplant recipients?

Methods
We performed this review according to Cochrane Handbook for Systematic Reviews of Interventions [23] and reported it by the Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) [24].

Data source and research strategy
We conducted a comprehensive search on Scopus and Medline (through PubMed) from 1980 until April 2019. The search was done using an arrangement of keywords and mesh terms that focused on self-management, empowerment, and participation as well as kidney transplantation. The combinations of keywords and MeSH terms in search strategies have been listed in Table 1.

Eligibility criteria
The factors used to determine the inclusion criteria were as follows: population, intervention, comparator, outcomes, and study design (PICOS). An IT-based system is defined as an automated system, without direct human intervention. To be included in the present study, the retrieved articles had to meet the following inclusion criteria: (1) IT-based interventions with an automated function, (2) interventions involving some type of ITbased tools to enable self-management, including smart phones, tablets, and computers, (3) published studies in scientific journals, (4) publication dated in the 1980-2017 period, (5) English language, (6) studies performed as a randomized clinical trial, which entails a control group which represents the standard care, without an ITbased intervention, and (7) studies performed on kidney transplant recipients.
The following articles were excluded from the study: (1) studies involving direct human interventions (e.g. nonautomatic phone calls, Short Message Service (SMS) and video representation systems).

Data extraction
After article retrieval, in the first phase of selection the articles' title and abstract was used as the basis to decide whether they fit the research question domain. In the second phase, selection was based on the inclusion criteria as described above, extracting studies that used ITbased solutions for self-management of kidney transplant patients. Following these two phases of article selection, the full text of the remaining articles was further investigated, and the remaining articles were categorized based on the type of IT-based tools.
Following variables were extracted from included articles: characteristics of study sample, performance variables, method description, type of intervention, duration of study, as well as defined outcomes and results. The same authors (RG, FK) independently reviewed full text of selected articles to be included for qualitative synthesis. Disagreements were resolved by discussion with the third reviewer (SE).

Risk of bias
The Cochrane Collaboration's evaluation tool was used to assess the risk of bias in clinical trials [25]. This tool has been designed to assess risk of bias in term of following domains: selection bias (i.e. random sequence generation and allocation concealment), reporting bias (selective outcomes reporting), performance bias (i.e. blinding of participants and personnel), attrition bias (incomplete outcomes data), detection bias (blinding of outcome assessor), and etc.

Synthesis and analysis
We did not conduct a meta-analysis because of inconsistencies in measured interventions outcomes. The articles combination was based on the classification of: outcomes, the type of intervention, and its impact. The outcomes of the studies were categorized into two groups: clinical and care process outcomes. Clinical outcomes are used to quantify or describe the severity of the disease, such as blood pressure. Care process outcomes are associated with improving the quality of care and physicianpatient interaction [26]. We used the Sign test to assess the effect of proposed intervention in either direction (e.g. positive or negative) for clinical and care process outcomes.
Based on the type of IT-based tools used in the interventions, we classified the studies as follows: 1 Mobile-based tools (i.e. educational contents which were delivered via smart phone applications or SMS). 2 Wearable devices (i.e. hardware devices which can be used for automatic recording of physiological changes, such as Holter monitor devices). 3 Computer systems, which enable the patient to record and transmit data and through the internet. 4 Multi-component tools, which is a combination of more than one tool from the above mentioned tools [27][28][29].
We used the technology performance framework to classify IT-based systems according to their function. This means classification of these systems based on whether they could: • Inform: Delivery media (e.g. text, voice, photo, and video). • Instruct: Offer instructions to the user • Record: Capture data entered by the user • Display: Show or output data entered by the user • Guide: Deliver guidance based on user provided information • Remind/Alert: Provide alerts and reminders for specific tasks or at specific times to the user • Communicate: Provide communication path between the user/patient and health care providers [28][29][30] The impact of interventions was defined as: (1) positive (i.e. statistically significant) and (2) no effect (i.e. not statistically significant). Figure 1 (Additional file 1) demonstrates the study selection process. A total of 2392 records (1170 from PubMed and 1222 from Scopus) were retrieved. After deduplication, 1813 unique studies remained. After evaluation of titles and abstracts, based on the inclusion and exclusion criteria, 36 papers were selected for full-text evaluation, among which 28 papers were excluded, and 8 papers were selected to be included in the study. Table 1 shows the general characteristics of the selected studies. In fulltext evaluation stage, 8 studies were selected, all published in English. The oldest study was published in 2010 [31], and the latest one was published in 2018 [32].

Study selection
Out of these 8 studies, five studies were conducted in the United States [3,[33][34][35][36], two studies were conducted in Germany [22,31] and one study was conducted in Canada [32]. Two studies included the patients on the waiting list of kidney transplant and six studies included kidney transplant recipients. The average sample size was 111 , and the median follow-up duration was 3 months (2 weeks-24 months). Majority of studies examined more than one (7/8 to 87.5%) outcome.
Out of these 8 studies, six studies were performed after transplantation. Other two studies examined three care process outcomes.

Bias risk assessment
The quality evaluation results for the studies are shown in Fig. 2 (Additional file 2). About 50% of the studies clearly described the method used to generate the allocation sequence, and about 37.5% of the studies reported the allocation concealment. Less than 65% of the studies reported the method they used to deal with the incomplete data, and 50% of the studies provided insufficient information about the method they used to blind the outcome assessors. As observed in the studies, despite the presence of reporting protocol, the outcomes were reviewed based on the pre-specified and reported outcomes. This bias was low in 12.5% of the studies. The bias was also low in assessing the cause of participants'  [22].

The impact of interventions on outcome measures
As shown in Table 2 a total of 30 outcomes variables were evaluated in the studies (i.e. 24 care process and 6 clinical outcomes). About 60% of the studies showed statistically significant positive effect in favor of using the proposed interventions. In the other 12 outcomes (40%), no significant difference was observed between intervention and control groups. Majority of outcomes were evaluated in interventions after transplantation (27/30, 90%). Out of 30 outcomes, 16 process of care and 4 clinical outcomes were reanalyzed using the Sign test. It should be noted that one study was excluded from the analysis due to incomplete baseline data [22]. The sign test was performed for clinical and process of care outcomes and showed that IT-based interventions were significantly affected the process of care outcomes (p = 0.021). However, no effect was observed for the clinical outcomes (p = 0.62). Seven studies [3,[31][32][33][34][35][36] evaluated sixteen process of care outcomes from which four studies had poor methodological quality [3,31,35,36]. Three studies affected the outcome in favor of the intervention [32][33][34]. Four clinical outcomes were assessed in three study and all of them had poor methodological quality [3,31,35] ( Table 3).

Clinical outcomes
The clinical outcomes evaluated in these studies were Glomerular Filtration Rate (GFR) changes (2 study), systolic blood pressure and diastolic blood pressure (1 study), tacrolimus whole-blood level (1 study) and acute rejection rate (1 study). Overall, the impact of IT-based interventions on clinical outcomes was significantly positive in 1 of the studies (33%), while in the other two studies (67%) there was no significant difference between the control and intervention groups.
In one study, the impact of computer-based patient education intervention on GFR changes was evaluated, having no significant effect [31]. In a study by McGillicuddy et al., the effect of an mHealth system (a blood pressure monitoring device) was evaluated on blood pressure, which also found a significant positive effect [3]. Another study evaluated the effect of IT-based interventions on the amount of tacrolimus, showing no significant difference between the intervention and control groups [35]. In another study, the effect of telemonitoring and real-time video consultations with access to significant medical data was evaluated on rejection rate and GFR, having no significant effect [22]. All clinical outcomes (6/6, 100%) were evaluated in interventions after transplantation.

Care process outcomes
The care process outcomes evaluated in this study consisted of patient's knowledge (3 outcomes), IRB and IRK (1 outcome), recognition of personal skin cancer risk (1 outcome), willingness to change sun protection (1 outcome), sun-protection use (1 outcome), daily hours outdoors (1 outcome), medication adherence (4 outcome), willingness to accept increased risk donor kidney (1 outcome), unplanned admission rate (1 outcome), length of unplanned stay (1 outcome), unplanned inpatient care costs (1 outcome), rejection therapy initiation (1 outcome), ambulatory care visit rate (1 outcome), quality of life (1 outcome), return to employment (1 outcome), selfefficacy (1 outcome), skills (1 outcome), medication side effects (1 outcome), self-perceived general state of health (1 outcome).      Computer systems (2) Multiple components (7) Smartphones or PDA (9) Computer systems (6) Multiple components (5) Smartphones or PDA (1) All in all, the impact of IT-based interventions on care process outcomes was significantly positive in 16 of 24 (66%) outcomes. In three outcomes, the effect of intervention on medication adherence was significantly positive; in one study, the use of an m-Health system (BP monitoring device) and in the other study, the use of wireless pill bottles monitoring with customized reminders (including alarms, texts, telephone calls, and/or e-mails) were evaluated on tacrolimus adherence [3,35]. In another study, the effect of tele-monitoring and real-time video consultations with access to significant medical data were evaluated on immunosuppressive adherence, which was found to have a positive effect [22]. On the other hand, one other study showed interactive web-based sessions have no significant effect on medication adherence [32]. Also, in 3 studies that evaluated the impact of tablet, mobile, and website accessibility on knowledge enhancement, there was a significant difference between the intervention and control groups [33][34][35]. Also, in another study, the effect of computer-based education on IRK and IRB was reported to be significantly positive [31]. Also, in one study, recognition of personal skin cancer risk, the desire to change sun protection and sun-screen use significantly increased using mobile app interventions, and daily hours outdoors significantly decreased using an intervention through the tablet [35]. Also, in one study, the desire to accept IRD kidney significantly increased using mobile app interventions [34]. In one study, an intervention by remote tele-monitoring and real-time video consultations with access to significant medical data improved unplanned admission rate, length of unplanned stay, unplanned inpatient care costs, quality of life, and return to employment [22].
Another study evaluated the effect of interactive webbased sessions hosted by a virtual nurse on self-efficacy, skills, medication side effects, and self-perceived general state of health and found no significant effects [32].
About 88% (21/24) of the care process outcomes were assessed in IT-based interventions after transplantation. About 67% (14/21) of the care process outcomes were significantly improved after implementing IT-based interventions. On the other hand, 33% (7/21) of the care process outcomes were not significantly different between the control and intervention groups.
Two studies were performed before transplantation. These studies 12% (3/24) evaluated care process outcomes. One study assessed the effect of an educational website on pre-transplant knowledge and found positive effect [33]. One study showed significant effect of a mobile-web application on pre-transplant knowledge. However, the willingness to accept the increased risk of donor kidney was not significantly improved [34]. Table 4 shows a summary of classification of interventions based on the type of technology. Three studies evaluated the effect of smart phone interventions using mobile health, mobile web applications, and a tablet program. The functions of smart phones consist of informing, communicating, and instructing. Smart phones had significantly positive effect on 9 out of 10 outcomes and showed no effect on only one outcome [3,34,36]. Three studies assessed the effect of computerized system interventions using a computer-based educational program and website. The functions of computerized systems involved instructing, informing, and communicating. These studies showed that interventions positively improved 2 out of 8 evaluated outcomes and was evaluated ineffective in six outcomes [31][32][33]. Also, two studies evaluated the impact of multi-component technologies, including a wearable tool, accompanied by SMS and telephone calls, as well as remote tele-monitoring and real-time video consultations with access to significant medical data. Their functions include recording, displaying, informing, instructing and communicating. The effect of using multi-component interventions was assessed as positive on seven outcomes, while it had no effect on 5 outcomes [22,35].

Discussion
This systematic review abstracted the clinical trials which evaluated the effect of IT-based interventions on the selfmanagement outcomes among kidney transplant recipients. A total of 6 studies including 930 patients showed significant improvement on the self-management outcomes. Majority of studies reached statistically significant effects (about 50% on clinical outcomes and 88.8% on process outcomes). Majority of the IT-based interventions were performed after transplantation (75%). Following medias were used: smart phones, wearable devices, computer systems, and multi-component interventions. The positive effect of IT-based interventions on clinical outcomes among transplant recipients is in accordance with previous systematic reviews which included the IT-based interventions [37][38][39][40]. Therefore, it can be concluded that IT-based tools are a suitable type of intervention to control clinical outcomes in kidney transplant recipients. The result of sign test for clinical and process of care outcomes showed that IT-based interventions significantly affected the process of care outcomes. There have been many trends in the use of ITbased technologies to educate patients and come up with better therapeutic options for various types of disease. Educating patients on disease and treatment is an effective way to increase awareness and self-management; however, knowledge is considered as one of the care process outcomes, which has been emphasized in most articles under study [31,33,34,36]. In fact, knowledge is a key outcome for self-management in kidney transplant patients. Knowledge has the ability to empower patients by enhancing awareness [41]. In this study, the effect of IT-based interventions on kidney transplant recipients' knowledge was reported as positive, consistent with results of other studies [42][43][44][45].
Increasing knowledge about the disease is a crucial aspect of a patient's capability for drug management [43]. There is an established relationship between medication adherence and clinical outcomes, so that non-adherence to medication is closely linked with hospitalization and increased rate of mortality [44]. In this regard, IT-based systems have the potential to improve transplant knowledge and increase adherence to immunosuppressive medications, thus providing self-management improvement, which, in turn, leads to reduced transplant rejection and improved quality of life [46]. For instance, reminders can specifically be used to target and change unintentional forms of behavior in non-adherent patients taking medications, such as amnesia. Reminders can also be used to improve drug adherence in all age groups [47]. Other systematic reviews have also confirmed the positive impact of reminders on improving drug adherence [48]. While some studies have reported a positive impact for the use of IT-based interventions on medication adherence [49][50][51], other studies have reported the impact of IT-based interventions as ineffective [52].
In the present review, we investigated the effect of m-Health on care process outcomes and found it to be positive. Previously it has been reported that using m-health solutions with different forms of applicability provide tools that can improve clinical outcomes [53,54]. It is an established fact that m-health can be used to improve quality, monitoring, and study of health-related data. For instance, the use of personalized learning tools requires a more active involvement of patients in the self-management process [55]. Previous studies have also shown that m-health solutions can improve the symptoms of the disease using active self-management interventions [54,56].
Blood pressure is an important clinical biomarker in kidney transplant patients because of its associated complications. In our study, it was shown that using ITbased interventions, we can better monitor and control this clinical outcome. Other previous studies have also reported that remote monitoring systems are helpful in controlling hypertension [57][58][59].
Only one study was free from the risk of bias [32]. Limited information about allocation concealment, random sequence, incomplete outcomes data, and blinding outcome assessor were the top four that contributed to the low risk of bias score in included studies.
One of the strength points of the present study was our comprehensive search strategy that collected a large number of studies, thus reducing the prospects of dropping relevant articles. Because only randomized clinical trials were included in this study and other types of studies were excluded, there was a lower risk of bias and the quality of publications was thoroughly examined.
One of the limitations of this study was lack of accessibility to conference papers. Another limitation of this study was the presence of heterogeneity in reported outcomes, which made meta-analysis not feasible. Future studies should consider a larger sample size that can increase the generalizability of the study, which would, in turn, increase the effectiveness of the desired outcomes. Half of the studies were conducted over a short period of time (less than a month). The duration of studies should be in accordance with the defined outcomes. Moreover, the studies are needed to be improved in terms of reporting bias. In another words, randomized controlled trials (RCTs) are potentially associated with low risk of bias.

Conclusions
IT-based interventions such as m-Health, wearable devices, and computer systems can improve self-management in kidney transplant recipients (including clinical and care process outcomes). It is suggested that these interventions begin before kidney transplantation and continue thereafter.