This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Medical Informatics, is properly cited. The complete bibliographic information, a link to the original publication on http://medinform.jmir.org/, as well as this copyright and license information must be included.
Shared decision making (SDM) is important in achieving patient-centered care. SDM tools such as decision aids are intended to inform the patient. When used to assist in decision making between treatments, decision aids have been shown to reduce decisional conflict, increase ease of decision making, and increase modification of previous decisions.
The purpose of this systematic review is to assess the impact of computerized decision aids on patient-centered outcomes related to SDM for seriously ill patients.
PubMed and Scopus databases were searched to identify randomized controlled trials (RCTs) that assessed the impact of computerized decision aids on patient-centered outcomes and SDM in serious illness. Six RCTs were identified and data were extracted on study population, design, and results. Risk of bias was assessed by a modified Cochrane Risk of Bias Tool for Quality Assessment of Randomized Controlled Trials.
Six RCTs tested decision tools in varying serious illnesses. Three studies compared different computerized decision aids against each other and a control. All but one study demonstrated improvement in at least one patient-centered outcome. Computerized decision tools may reduce unnecessary treatment in patients with low disease severity in comparison with informational pamphlets. Additionally, electronic health record (EHR) portals may provide the opportunity to manage care from the home for individuals affected by illness. The quality of decision aids is of great importance. Furthermore, satisfaction with the use of tools is associated with increased patient satisfaction and reduced decisional conflict. Finally, patients may benefit from computerized decision tools without the need for increased physician involvement.
Most computerized decision aids improved at least one patient-centered outcome. All RCTs identified were at a High Risk of Bias or Unclear Risk of Bias. Effort should be made to improve the quality of RCTs testing SDM aids in serious illness.
Shared decision making (SDM) is important in achieving patient-centered care, as it involves both the patient and the health care provider in medical decision making [
SDM tools such as decision aids are intended to inform the patients with regard to the risks, benefits, and trade-offs associated with a decision [
Computerized decision aids can offer personalized evidence-based care, and if they are presented in an SDM capacity they can result in treatment decisions that respect the autonomy and preferences of the patient. Additionally, technological advances that use and process electronic health record (EHR) data may allow for the development of large-scale, low-cost assessments that can improve patient goals [
A recent systematic review by Austin et al [
This systematic review builds on the work of Austin et al [
The preferred reporting items for systematic reviews and meta-analyses (PRISMA) checklist for systematic reviews was followed for this review. The study was not registered with the International Prospective Register of Systematic Reviews (PROSPERO), and therefore, registration information is not included.
PubMed and Scopus databases were searched in February 2016. The search was conducted without a limitation on the year of publication. The search strategy terms were based on the terms used in the systematic review of similar topic by Austin et al [
The search terms utilized in the PubMed database were as follows:
The above search was then modified for the Scopus database:
Papers extracted from the search results mentioned SDM tools/aids, communication tools/aids, or SDM in relation to an illness or disease in the title or abstract. The abstracts and/or full text were then reviewed; papers were included if the study design was determined to be an RCT. Papers that assessed the use of noncomputerized tools or aids such as videos or pamphlets were excluded as the purpose of this review was to consider computerized decision tools. Tool formats included were Web-based, EHR portals, or computerized decision support software. Included RCTs had to discuss the use of computerized decision aids in serious illness as defined in the introduction. Finally, the paper had to discuss the tool in relation to aspects of SDM such as reducing decisional conflict and increasing knowledge. Tools were included if they were for the use of patients and/or family of patients. The patient population considered included both adults and children living with serious illness.
The references in the selected papers were hand-searched for relevant papers. Data from the final papers were manually extracted. Only published papers and papers in English were included in the study. The selection of papers was completed by one investigator.
Papers were graded on quality using a Cochrane Risk of Bias Tool (Modified) for Quality Assessment of Randomized Controlled Trials. The quality assessment included the following study validity domains: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other bias. Studies were assessed as either High, Low, or Unclear Risk of Bias. Identified problems in one domain would result in the study being labeled as “High Risk of Bias.” Assessment of the quality of the selected papers was completed by 2 investigators. In case of a disagreement between the 2 investigators completing quality assessment, a third investigator was consulted.
Study characteristics of all included RCTs were described according to PRISMA systematic review guidelines. All patient-centered outcomes in relation to SDM or communication were described, regardless of whether they differed significantly from the control. Patient-centered outcomes extracted from studies varied and included satisfaction with decision, decisional conflict, clinical outcomes, knowledge, preparation for decision making, emotional well-being, perceived involvement in medical decision making, patient expectations, satisfaction with physician discussion, parental activation, and number of school or work days missed. As
A total of six papers describing RCTs of SDM tools for serious illness were selected and reviewed (
The results suggest that computerized decision aids may be used for various types of serious illnesses in a variety of different health care settings to assist both patients and clinicians in decision making. Generally, the selected RCTs demonstrated that computerized decision aids were able to reduce decisional conflict [
Each of the formats of the computerized decision aids included common features among them. The Web-based decision tools commonly used surveys or questionnaires to ascertain patient preferences, which were then used to guide patient-physician communication or to provide treatment options. The EHR portal decision tool featured the ability for patients to track relevant information and provided educational content, both of which were ultimately used to guide treatment plans. Estimates of treatment efficacy and prognosis were common in interactive computer applications. Of the six selected RCTs, Web-based decision tools were described by Meropol et al [
In a single-blind RCT, Meropol et al [
Literature search and selection.
There were no statistically significant differences between the two different intervention arms on any of the satisfaction or decisional conflict responses; the summary report for the physician did not improve outcomes. Intervention arms were combined and analyzed against the control arm. Participants assigned to intervention groups had higher levels of satisfaction with discussions about the format of physician communications and quality of life issues but did not differ in satisfaction of discussion regarding diagnosis/prognosis, treatment options, or support community services. Those in the intervention arms found that CONNECT made it easier to reach treatment decisions and were more satisfied with their treatment choice. Participants in the intervention groups had decreased expectations of severe side effects with standard or experimental therapy. The CONNECT intervention was associated with increased satisfaction with overall communication in those with postsecondary education. Additionally, patients in the intervention arm reporting a lower baseline quality of life had greater satisfaction with overall communication.
The study was limited by a racially and ethnically homogenous sample population that was mostly gathered from large cancer centers. Furthermore, the eligibility criteria limited the study to include only those with personal Internet access or those who could arrive early to their appointments to access computers on-site. Additionally, patients in the control groups were directed to the National Cancer Institute website, where extensive searching by the patients may result in a reduced difference between groups. Furthermore, the merging of the intervention groups may place the study at risk of reporting bias. Using the modified Cochrane Risk of Bias Tool for Quality Assessment of Randomized Controlled Trials, the RCT by Meropol et al [
An RCT by van der Krieke et al [
Perceived involvement in medical decision making did not differ from patients in the control condition. There were no differences in self-reported satisfaction with care between study arms. However, within the intervention group, those who received the allocated intervention reported lower satisfaction with care in comparison with those who did not receive the intervention.
The study demonstrated a low response rate (29.2%, 73/250) and a moderate participation rate. Furthermore, the study protocol was weakly implemented; not all participants in the intervention group were offered the possibility to use the decision aid, and treatment evaluation meetings where the SDM process would have been used to guide treatment plans did not always occur. The authors do not provide sufficient information regarding the blinding process, if any, that was implemented in the study; therefore, there is unclear risk of selection, performance, and detection bias. The study by van der Krieke et al [
A Web-based, tailored, interactive health communication application for patients with T2D or CLBP was tested in an RCT by Weymann et al [
Intention-to-treat analysis, which used the baseline data, found no statistically significant differences between the groups; however, there was a significant difference between T2D and CLBP users, indicating higher knowledge scores in the T2D group. Conversely, sensitivity analysis, which used data from the available cases, found that participants using the tailored system displayed more knowledge immediately after the first visit than those in the control group. Additionally, those in the intervention group had more emotional well-being as identified by a subscale of a patient empowerment scale at the 3-month follow-up. Sensitivity analysis did not result in significant differences between the intervention and control groups in decisional conflict and preparation for decision making.
The sample population was only limited to those with personal Internet access, which may not be representative of the general population. Additionally, the study did not assess outcome criteria at baseline or address potential confounders, both of which make it unclear whether any observable differences were a result of the intervention or other factors. The measure used to assess T2D/CLBP knowledge, a primary outcome of the trial, was also not validated. Moreover, despite blinding of the participants, the use of the dialogue format may have allowed participants to identify the intervention. The study by Weymann et al [
An RCT by Fiks et al [
The authors reported no statistically significant differences between the control and intervention groups’ satisfaction with asthma care or medication receipt, but data were not made available in the study report. There was no effect on parental knowledge, skills, and confidence. Parents in the intervention group had a significant decrease in the number of days of work missed in comparison with the controls. Analysis indicated an improvement of the frequency of asthma flares in the intervention group compared with the control group. There were no differences in quality of life measurements between the two groups; however, compared with the control group, families of intervention group reported fewer emergency department visits and hospitalization over 6 months. Portal use was also found to be greater in parents of children with moderate to severe asthma than those whose children had mild persistent asthma.
As the participants were recruited based on referrals by physicians or EHR rosters, the sample is considered a convenience sample and its representativeness is unclear. Also, because of the small sample size, randomization did not result in a balance between intervention and control groups in terms of asthma severity. The inadequate randomization of participants places the study at a risk of selection bias, and therefore, the study by Fiks et al [
An RCT by Hochlehnert et al [
There was no significant difference in satisfaction with decision or decisional conflict, as well as assessment of information tool between the two groups. The two groups were merged for analysis, and it was found that those who were satisfied with the information presented in the tool experienced more satisfaction with their decision and experienced less decisional conflict. Furthermore, those who perceived the tool to be useful in a general practitioner’s office and were satisfied with introduction of the tool (ie, training) were more likely to be satisfied with their decision.
The authors do not provide sufficient information regarding the blinding process that was implemented in this trial; therefore, the risk of performance and detection bias is unclear. The study by Hochlehnert et al [
An RCT by Peele et al [
Women who received the decision aid were significantly less likely to choose adjuvant therapy than those in the control group; one-third fewer women in the intervention group received adjuvant therapy than their counterparts in the control group. The impact of the decision aid based on tumor severity found that the participants in the intervention group with low tumor severity rejected adjuvant therapy significantly more often than the participants in the control group. Generally, women with higher tumor severity, younger women, and women with a university-based physician were more likely to choose adjuvant therapy.
Neither patients nor clinicians were blinded in this study, indicating risk for performance and detection bias. A higher proportion of university-based physicians were randomized into the intervention group, which places the study at a risk of selection bias as well. The study by Peele et al [
Study results from the six RCTs discussed in the Results section demonstrate that computerized decision aids have the potential to improve patient-centered outcomes. Furthermore, decision aids have differing impacts on various patient-centered outcomes that can possibly be attributed to tool design, user characteristics, or type of disease. Coincidentally, in this review, each of the selected RCTs employed computerized decision aids in management of chronic illnesses, although this was not specified in the search strategy. Furthermore, the small number of studies that are included in this review also suggests that there is still much work to be done in this area. Of the six computerized decision aids discussed in this review, only the tool used by Hochlehnert et al [
Decisional conflict was addressed in four RCTs [
EHR portals that function as a decision support system for both patients and physicians present a unique opportunity to manage care from the home. The MyAsthma portal for pediatric asthma did not have an effect on quality of life measures but did result in decreased days of work missed by parents of pediatric patients and a reduction of asthma flares [
The information presented in a decision tool is of importance to achieving meaningful patient-centered outcomes. Hochlehnert et al [
It is possible that patients may benefit from decision tools without the need for specialized communications training or extra involvement of physicians. The computerized information tool for fibromyalgia patients was tested with and without consultation of a physician specially trained in facilitating SDM. There were no statistical differences between groups on any patient-centered outcomes, including decisional conflict or satisfaction with decision [
A computerized decision aid, such as Adjuvant!, can present the risks and benefits of treatments to the patients and allow them to consider their preferences and values when making treatment decisions. This may result in a reduction in therapies that are not in line with patient preferences or disease severity and, consequently, can reduce treatment cost: Adjuvant! demonstrated a reduction in adjuvant therapy, such as chemotherapy, in breast cancer patients and was effective at decreasing adjuvant treatment in patients with low tumor severity [
Only one study compared computerized decision aids against a nontechnological decision aid. Adjuvant! resulted in decreased use of adjuvant therapy in comparison with control group participants who received informational pamphlets about adjuvant therapy [
The tools discussed in this review are relatively simple from a technological perspective. There is potential for greater detail and personalization in SDM with the advent of more advanced decision support tools and the widespread of EHRs. For example, it has been suggested that dynamic clinical data mining can be used to provide real-time decision support. Search engine queries of a population database built on deidentified EHR would provide clinical data support using prior clinical cases, relevant statistics, scholarly resources, and protocols [
A limitation of this study was the quality of the RCTs selected for review. The RCTs included in this review were either at High Risk of Bias or Unclear Risk of Bias. Risk of bias should be considered when assessing the strength of evidence provided by the RCTs in this review. The literature search was also only limited to published papers and is therefore subject to publication bias. Furthermore, the search was only limited to PubMed and Scopus databases. Although these databases consist of an extensive amount of literature on the topic, the results may not have been representative of the entirety of the literature. Additionally, secondary search strategies were not performed. The study was also not registered in PROSPERO, which limits the study in terms of adhering to current best practices for systematic reviews. Finally, although quality assessment of the selected papers was completed by 2 investigators, the study is limited because of the fact that the selection of papers was completed by only a single investigator.
Most computerized decision aids improved at least one patient-centered outcome. The RCTs differed in patient outcomes measured and the efficacy of decision aids in improving the aspects of SDM. All RCTs identified were at High Risk of Bias or Unclear Risk of Bias according to a modified version of the Cochrane Risk of Bias Tool for Quality Assessment of Randomized Controlled Trials. Efforts should be made to improve the quality of RCTs testing SDM aids in serious illness.
chronic lower back pain
communication skills training
electronic health record
interaction health communication applications
preferred reporting items for systematic reviews and meta-analyses
International Prospective Register of Systematic Reviews
randomized controlled trials
shared decision making
type 2 diabetes
This work was supported by a Discovery Grant (RGPIN-2014-04743) from the Natural Sciences and Engineering Research Council of Canada (NSERC).
None declared.
Study description, effect of decision tools on patient-centered outcomes, and risk of bias.