This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
While Web-based interventions have been shown to assist a wide range of patients successfully in managing their illness, few studies have examined the relative contribution of different Web-based components to improve outcomes. Further efficacy trials are needed to test the effects of Web support when offered as a part of routine care.
Our aim was to compare in regular care the effects of (1) an Internet-based patient provider communication service (IPPC), (2) WebChoice, a Web-based illness management system for breast cancer patients (IPPC included), and (3) usual care on symptom distress, anxiety, depression, (primary outcomes), and self-efficacy (secondary outcome). This study reports preliminary findings from 6 months’ follow-up data in a 12-month trial.
We recruited 167 patients recently diagnosed with breast cancer and undergoing treatment from three Norwegian hospitals. The nurse-administered IPPC allowed patients to send secure e-messages to and receive e-messages from health care personnel at the hospital where they were treated. In addition to the IPPC, WebChoice contains components for symptom monitoring, tailored information and self-management support, a diary, and communication with other patients. A total of 20 care providers (11 nurses, 6 physicians, and 3 social workers) were trained to answer questions from patients. Outcomes were measured with questionnaires at study entry and at study months 2, 4, and 6. Linear mixed models for repeated measures were fitted to compare effects on outcomes over time.
Patients were randomly assigned to the WebChoice group (n=64), the IPPC group (n=45), or the usual care group (n=58). Response rates to questionnaires were 73.7% (123/167) at 2 months, 65.9 (110/167) at 4 months, and 62.3% (104/167) at 6 months. Attrition was similar in all study groups. Among those with access to WebChoice, 64% (41/64) logged on more than once and 39% (25/64) sent e-messages to care providers. In the IPPC group, 40% (18/45) sent e-messages. Linear mixed models analyses revealed that the WebChoice group reported significantly lower symptom distress (mean difference 0.16, 95% CI 0.06-0.25,
In spite of practice variations and moderate use of the interventions, our results suggest that offering Web support as part of regular care can be a powerful tool to help patients manage their illness. Our finding that a nurse-administered IPPC alone can significantly reduce depression is particularly promising. However, the multicomponent intervention WebChoice had additional positive effects.
Clinicaltrials.gov:NCT00971009; http://clinicaltrials.gov/show/NCT00971009 (Archived by WebCite at http://www.webcitation.org/6USKezP0Y).
The number of Web-based support systems to enhance self-management for people living with health conditions has increased rapidly in the last decade, and such interventions have been shown to assist a wide range of patients [
However, it can be difficult to distinguish which components of Web-based support systems are most beneficial for patients, and little is known about the relative contribution of different components [
One component often offered as part of multicomponent systems, or as a standalone service, is e-messages. Several studies report benefits from using Internet-based patient-provider communication services (IPPC) for communication between patients and health care providers in terms of assisting patients in managing illness and improving health outcomes [
WebChoice is a Web-based illness management support system based on patient-centered principles and designed to support cancer patients in self-management of their illness, independent of location and time [
As several studies show benefits to Web-based support systems, it is timely to examine the relative contribution of different components of these multicomponent support systems aiming to improve selected outcomes. As cancer patients experience their illness primarily through symptoms, reduction of symptom distress, anxiety, and depression are important indicators of the success of illness management support. Furthermore, there is a need to test the effects of Web-based support as a part of regular care.
The aims of the study were therefore to test and compare the effects of (1) an IPPC, (2) the multicomponent WebChoice intervention (including an IPPC), and (3) usual care (control group) on symptom distress, anxiety, and depression (primary outcomes), as well as self-efficacy (secondary outcome) after 6 months of follow-up. In addition, explorative sub-analyses were performed to detect whether the outcomes were associated with the actual use of the interventions. We hypothesized that the WebChoice group compared with the usual care group would have better outcomes than the IPPC group compared with the usual care group. We also hypothesized that both groups would have better outcomes than the usual care group, on both primary and secondary outcomes.
We conducted an RCT with three groups: two intervention groups (the IPPC service and WebChoice) and a usual care group (clinical trial NCT00971009). Due to slower recruitment than anticipated, we had to stop inclusion after 200 consenting participants, before the calculated sample was obtained. The current paper reports on 167 patients for whom 6-month follow-up data were available at the time this paper was written. These participants were recruited between May 2010 and September 2012.
Inclusion criteria were recent diagnosis of breast cancer treated with surgery, or under treatment with radiation, chemotherapy, hormone therapy, or combinations of those (maximum 12 months after surgery), age over 18 years, able to write/read and speak Norwegian, having access to the Internet at home, and having a public key infrastructure (PKI) solution for secure system access.
Study participants were recruited from three hospitals in Norway—one university hospital and two regional hospitals—at breast diagnostic centers or the ambulatory chemotherapy, radiation, and surgical units. Participants did not receive any incentives for participating in the study.
See
Eligible patients scheduled for surgery or coming in for checkups after surgery or treatment were identified by the study nurses at the hospitals and provided with information about the study. Upon patients’ arrival at the clinic, the study nurses met the patients, provided brief information about the study, and asked if they were interested in participating. If the patients agreed, the nurse informed them about the study’s purpose and procedures and asked for written informed consent. Consenting patients completed baseline questionnaires before randomization.
After completion of baseline questionnaires, patients were randomized according to a pre-defined automated computerized block randomization, with a block size of 42 stratified by site. Due to the content of the interventions, patients could not be blinded to which arm they were randomized.
Patients randomized to the usual care group were followed up as usual at the hospitals where they were treated. Patients who were randomized into the IPPC or WebChoice groups were informed and instructed in the use of the IPPC or WebChoice. They received a printed user manual with instructions for use, how to log on to the system, and an address and phone number to contact for help if needed. The only in-person information given was instructions on how to access the site and how to connect with the study support service if questions occurred. The study nurses showed them where in the user manual they could go to find information about how to access the site and how to connect with the study support service if needed. In addition, the participants were informed that they could use the IPPC or any component of WebChoice as much or as little as they liked and that using the system was entirely voluntary. The IPPC component was the same for both intervention groups.
After being informed about group assignment, the patients were given access to the interventions the same day. They received an automatic welcome message when the system was ready to use. There was an option to be notified by text message or regular email when new messages appeared in the system. Most participants wanted this notification. All participants were sent questionnaires by postal mail at 2, 4, and 6 months after enrolling.
In total, 20 care providers answered questions from patients: a dedicated group of expert nurses (n=11) and physicians (n=6) in breast cancer care, and social workers (n=3) at the hospital where the patients were treated. They were thoroughly trained in administering the IPPC, technically as well as in codes of conduct for online communication with patients. There was a clear schedule for who was responsible for answering patients’ messages. The nurses were frontline and received all messages first. If necessary, they could forward the message to other care providers. If considered important, information from e-messages could be copied into the medical record and made available for other health care providers. When new questions arrived in the system, the recipient was notified through the hospital’s email system or by text message. The same providers answered e-messages from both the IPPC and WebChoice groups using the same interface. However, they were not entirely blinded to the intervention group assignment because this was sometimes disclosed by patients through the messages. The health care providers had no access to details about how patients used other components of WebChoice. They did not receive any reimbursement or additional dedicated time for answering secure e-messages from the participants.
The study was approved by the Regional Committee for Medical and Health Research Ethics and the Data Security Inspectorate in Norway. Written informed consent was obtained from all participants. All data were submitted to a secure server using an encrypted connection. Patients and health care personnel were authenticated using a public key solution that is currently used by Norwegian banks as a security platform. This means the users’ logon procedure is the same whether they log on to their online bank or to IPPC/WebChoice. Thus, patients did not need to learn a new procedure.
The system experienced a 9-day period of downtime during the first year due to technical problems at the hospital server that hosted the application. No changes were made to the interventions during the trial period apart from fixing minor bugs. The interventions could be used from different Internet browsers and were independent of Internet connection speed.
The IPPC is a further development of the IPPC component described in a previous study of WebChoice [
Message flow between patients and health care providers.
WebChoice was developed in close cooperation with users and health care personnel [
An assessment component where patients could monitor their symptoms, problems, and priorities for support along physical, functional, and psychosocial dimensions. From a predefined list, patients could choose symptoms and problems they were experiencing, rate the burden of these, and indicate where they needed help. This information could be used to monitor improvement/deterioration of the condition, indicate when to alert health care personnel, prepare for a hospital/physician consultation, improve patient-provider communication, or obtain immediate access to the self-management advice components described below.
An advice component provided illness self-management support. The patients’ self-reported symptoms triggered the display of appropriate self-management activities that patients could choose from to relieve symptoms and problems. The component could also be used without finishing an assessment first. Each choice contained an explanation of what the activity was, how to perform it, potential risks, side effects, contraindications, when to contact a physician, levels of evidence, references to the source of the evidence, and links to other reliable websites for related information. The advice component was updated once a year.
An information component where patients had access to other reliable Web sources in Norwegian and English, such as information about tests, treatments and potential side effects, lifestyle suggestions, and information about patients’ legal rights. External links were automatically checked every fourth week to ensure they were still active.
A communication component for sharing experiences with other patients. Patients could participate in an online forum group discussion that allowed them to exchange messages anonymously with other patients or use a blog. The forum and blog were monitored by nurses at our research center. The nurses did not participate in the forum or blog but answered two postings in the forum not answered by the other participants.
An electronic diary where patients could keep personal notes.
Screenshot of the WebChoice overview page.
The primary outcomes were symptom distress, anxiety, and depression. The secondary outcome was self-efficacy. All outcomes were measured at baseline, 2, 4, and 6 months thorough self-assessed questionnaires sent to participants by postal mail.
Patients provided information on age, marital status, level of education, employment status, income, and the use of Internet services.
The time of diagnosis and stage of disease were obtained from the medical record. Based on the patients’ tumor (T), node (N), and metastasis (M) classification at the time of diagnosis, the stage of disease was classified into 5 stages (0=ductal carcinoma in situ, to 4=advanced-stage disease) using the TNM Classification of Malignant Tumours of the Union for International Cancer Control guidelines [
Patients completed the Self-Administered Comorbidity Questionnaire (SCQ-19), which evaluated the number of, treatments for, and functional impact of health problems. It includes 16 common comorbidities and three optional conditions [
Symptom distress was measured by using the 32-item Memorial Symptom Assessment Scale (MSAS) [
Anxiety and depression were measured with the Hospital Anxiety and Depression Scale (HADS) [
Self-efficacy was measured with the Cancer Behavioral Inventory (CBI) version 2.0 [
Data on system use were extracted from the user logs on the server. Information was collected on how many times the users had logged on and which components of WebChoice were accessed or used actively.
Data on baseline characteristics are presented as medians and ranges for continuous variables and as proportions with percentages for categorical data. Differences between users and non-users were analyzed using the chi-square test for pairs of categorical variables. The Mann-Whitney-Wilcoxon test was used for continuous data with skewed distributions.
For analysis of between-group differences in symptom distress, anxiety, and depression (primary outcomes) and self-efficacy (secondary outcome), linear mixed models (LMM) for repeated measures were fitted. A diagonal covariance structure was used to model dependencies among measurements on the same individual at different time points. Models for each outcome consisted of 3 effects: measurement occasion (time), interventions (WebChoice, IPPC, usual care), and the interaction of time and intervention. All measured time points of the outcome variables are considered and the LMM approach therefore adjusts for baseline differences. To test whether potential confounders impacted the results, LMM adjusted for variables such as site, age, marital status, education, time since diagnosis, stage of disease, and comorbidity were fitted. Compared to the unadjusted models, these adjusted models revealed even larger differences in favor of the intervention groups compared to the usual care group. Taking the limited sample size into account and aiming to avoid over fitting, only the results from the unadjusted models are presented. As no statistically significant differences were observed between the study groups on demographic and disease-related factors at baseline, these models were not further adjusted for the possible confounders. The authors are aware that this might underestimate the true differences between the groups. Analyses of primary and secondary outcomes were conducted on an intention-to-treat basis, including all participants in each group, independent of whether they were users or non-users of the interventions. The model parameters are estimated using the classical maximum likelihood approach. No imputation of missing data was necessary or performed, as the LMM uses all data available to estimate the covariance matrix and model the dependencies. The results are presented as
In addition, explorative sub-analyses were performed to detect whether the outcomes were associated with the actual use of the interventions. LMM for repeated measures were fitted. Models for each outcome were fitted with three factors: measurement occasion (time), interventions (user/non-user of WebChoice and IPPC), and the interaction of time and intervention. Age was added as a covariate because age is known to be associated with use of Web-based tools [
Analyses were carried out using SPSS version 18.0.
The trial flow chart (
Flow of participants throughout the trial.
There were no statistically significant differences in demographics, disease-related factors, or outcome measures between participants in the three groups at baseline (
Baseline demographic and illness characteristics (n=167).
Characteristics | WebChoice (n=64) | IPPC (n=45) | Usual care (n=58) |
|
|
Age in years, median (range) | 51 (37-79) | 50 (31-66) | 53 (36-69) | .21 | |
|
.63 | ||||
|
Married/cohabitating | 52 (82) | 34 (76) | 43 (77) |
|
|
Single/divorced | 11 (18) | 11 (24) | 13 (23) |
|
|
.29 | ||||
|
Elementary/high school | 23 (37) | 22 (49) | 30 (54) |
|
|
University/college ≤4 years | 24 (39) | 14 (31) | 12 (21) |
|
|
University/college >4 years | 15 (24) | 9 (20) | 14 (25) |
|
|
.46 | ||||
|
<400,000 | 8 (13) | 12 (27) | 10 (18) |
|
|
400,000 to 600,000 | 14 (23) | 9 (20) | 18 (32) |
|
|
600,000 to 800,000 | 19 (31) | 11 (24) | 13 (23) |
|
|
>800,000 | 21 (34) | 13 (29) | 15 (27) |
|
|
.54 | ||||
|
Full-time/part-time work | 21 (33) | 18 (41) | 14 (26) |
|
|
Sick leave/disability benefits | 34 (54) | 22 (50) | 32 (58) |
|
|
Retired/other | 8 (13 ) | 4 (9) | 9 (16) |
|
|
.53 | ||||
|
0 | 2 (3) | 1 (2) | 5 (9) |
|
|
1 | 24 (38) | 21 (47) | 26 (45) |
|
|
2 | 32 (50) | 20 (47) | 24 (41) |
|
|
3 | 6 (9) | 2 (4) | 3 (5) |
|
|
|||||
|
Symptom distress | .47 (.03-1.49) | .52 (.03-1.76) | .59 (.05-1.86) | .36 |
|
Anxiety | 4.0 (0-16) | 5.0 (0-17) | 4.0 (0-15) | .38 |
|
Depression | 1.5 (0-11) | 2.0 (0-10) | 2.0 (0-14) | .37 |
|
Self-efficacy | 247 (110-297) | 214 (101-297) | 238 (102-297) | .11 |
Months since diagnosis, median (range) | 1 (0-10) | 0 (0-9) | 1 (0-10) | .16 | |
Comorbidity, median (range) | 2 (0-10) | 2 (0-16) | 2.0 (0-13) | .51 | |
|
|||||
|
> 1x/week | 57 (91) | 37 (82) | 53 (95) | .09 |
|
< 1x/week | 2 (3) | 6 (13) | 3 (5) |
|
|
< 1x/month | 4 (6) | 2 (4) | 0 (0) |
|
|
.28 | ||||
|
> 1x/week | 54 (87) | 36 (80) | 43 (78) |
|
|
< 1x/week | 3 (5) | 2 (4) | 7 (13) |
|
|
< 1x/month | 5 (8) | 7 (16) | 5 (9) |
|
|
.67 | ||||
|
> 1x/week | 28 (45) | 18 (40) | 20 (36) |
|
|
< 1x/week | 7 (11) | 9 (20) | 8 (15) |
|
|
Never | 27 (44) | 18 (40) | 27 (49) |
|
aNOK=Norwegian kroner: 400,000≈$US 67,000; 600,000≈$US 100,000.
The WebChoice group tended to score higher than the usual care group on self-efficacy (secondary outcome) over time (mean difference 8.81, 95% CI -0.92 to 18.53,
Effect of IPPC (n=45) and WebChoice (n=64) compared with the usual care group (n=58) on outcome variables: summary of basic mixed models.
|
WebChoice | IPPC | ||||
|
Mean diffa | 95% CI |
|
Mean diffa | 95% CI |
|
MSAS total | -0.16 | -0.25 to -0.06 | .001 | -0.07 | -0.18 to 0.04 | .21 |
HADS-A | -0.79 | -1.49 to -0.09 | .03 | -0.14 | -0.93 to 0.64 | .72 |
HADS-D | -0.61 | -1.18 to -0.05 | .03 | -0.69 | -1.32 to -0.05 | .03 |
CBI | 8.81 | -0.92 to 18.53 | .08 | -4.89 | -15.90 to 6.12 | .38 |
aBased on estimated marginal means. Analyses adjusted for baseline scores.
Estimated marginal means of (A) symptom distress (MSAS Total), (B) anxiety (HADS Anxiety), (C) depression (HADS Depression), and (D) self-efficacy (CBI Total) for the usual care group (n=58), the IPPC group (n=45), and the WebChoice group (n=64).
In accordance with the CONSORT-EHEALTH checklist [
The IPPC messages were mainly answered by nurses within their regular working hours. Of the 153 messages sent by the respondents, 22% (33/153) were passed on to and answered by physicians. Only one message was passed on to social workers. Patient questions posed through the IPPC were always answered via secure e-messages. In a few instances, the nurses informed the patient that they could contact her by phone to discuss issues voiced in the message. Time spent on answering messages was not measured. However, there were no indications that the nurses or physician felt the e-message answering task had been too time consuming for them during the study.
There were no differences between users and non-users in either intervention group on any demographic or disease-related variables.
Usage of components in WebChoice (n=41) and IPPC (n=18) over 6 months of access.
|
Users WebChoice (n=41) | Users IPPC (n=18) | ||||||||
Components | Times accessed | Users who accessed at least once | Times accessed | Users who accessed at least once | ||||||
|
Median | IQRa | Range | n | % | Median | IQRa | Range | n | % |
Total visits | 7 | 11 | 2-41 | 41 | 100 | 9.5 | 10 | 1-17 | 18 | 100 |
E-messages sent | 1 | 3 | 0-9 | 25 | 61 | 3.5 | 4 | 1-7 | 18 | 100 |
Total e-messages visits | 4 | 7 | 0-25 | 39 | 95 | 12 | 12 | 2-24 | 18 | 100 |
Assessments | 1 | 2 | 0-7 | 30 | 77 |
|
|
|
|
|
Assessment visits | 5 | 6 | 0-25 | 40 | 98 |
|
|
|
|
|
Advice views | 5 | 10 | 0-62 | 29 | 71 |
|
|
|
|
|
Information section visits | 2 | 3 | 0-28 | 34 | 83 |
|
|
|
|
|
Posts in forum | 0 | 0 | 0-3 | 8 | 20 |
|
|
|
|
|
Forum visits | 4 | 4 | 0-25 | 38 | 83 |
|
|
|
|
|
Posts in blog | 0 | 2 | 0-9 | 15 | 37 |
|
|
|
|
|
View of others’ blogs | 14 | 19 | 0-58 | 37 | 90 |
|
|
|
|
|
Diary notes | 0 | 3 | 0-14 | 19 | 46 |
|
|
|
|
|
Diary visits | 3 | 6 | 0-26 | 35 | 85 |
|
|
|
|
|
aIQR=interquartile range.
No differences were detected between users and non-users of WebChoice or IPPC on symptom distress, depression, or self-efficacy (data not shown). The users of the IPPC had significantly lower scores on anxiety compared with the non-users (mean difference -1.28, 95% CI -2.54 to -0.01,
The current effectiveness study demonstrates that access to the multicomponent Web-based support system WebChoice for 6 months, among women with breast cancer within the first year after diagnosis, reduced symptom distress and levels of anxiety and depression scores. A tendency towards increased self-efficacy could also be detected for the WebChoice group. This is promising given three diverse practice settings in regular care. Also noteworthy is the finding that access to an IPPC alone reduced depression scores. These results support the hypothesis of the WebChoice group having better outcomes than the IPPC group in symptom distress and anxiety, compared to the usual care group. The IPPC group had a similar effect on reduction of depression as WebChoice, however, and the WebChoice group did not have better self-efficacy than the IPPC group compared to usual care.
Our findings are in line with previous research showing that Web-based support systems in cancer populations can decrease depression and anxiety scores [
One possible explanation for our significant results might be that both WebChoice and IPPC might meet unmet needs reported by cancer patients, such as needs within the communication, information, psychosocial, psychological, and supportive care domains, generally highest during the treatment phase [
Our results indicating that the IPPC reduced depression scores are especially promising, as depression is one of the most debilitating symptoms people can have and is highly prevalent among cancer patients [
The IPPC’s ability to reduce depression scores might relate to indications that patients with higher scores of depression, in addition to higher symptom distress and low social support, are high users of the IPPC service in WebChoice [
Whether the differences in effects of the IPPC and WebChoice are related to the additional features contained in WebChoice is not clear. The IPPC feature might be used differently in the two groups, and as such we were not able to fully disentangle the effect of the IPPC feature. The patients would have had to be allocated to separate conditions in order to test different features. However, as in the IPPC group where 40% (18/45) sent e-messages, 39 % (25/64) in the WebChoice group sent e-messages, suggesting that interest in using this component is the same, independent of other features available.
Use was not connected to the observed effects on symptom distress and depression in the current study. Similar results were observed in the previous study of WebChoice [
Several limitations need to be addressed. The first limitation concerns the small sample size. A larger sample would have increased the validity of the study but would also have prolonged the time needed for recruitment. In the recruitment process, most of the potential participants were approached. However, only a third of those approached were included. Lack of access to Internet, the most frequent reason for not meeting the inclusion criteria, was reported by 19% (98/522) of those approached. Among those who were eligible, a frequent reason given for declining participation was that patients judged their computer and Internet skills as poor. One way to increase the participation rates might be to offer a demonstration of the interventions at the time of inclusion.
A smaller sample size than initially calculated and the attrition rate during the study reduced statistical power for our analyses. Because we had to stop inclusion of participants before the a priori calculated sample was obtained, block randomization led to different sample sizes in the three groups, with the least participants in the IPPC group. The project was also subject to high attrition during the study, which is not uncommon in studies of eHealth interventions [
Another limitation relates to low use of the interventions. The analyses of the intervention groups compared to the usual care group therefore compare the effects of a little used intervention. However, as the effects were detected through intention to treat analysis, the effect might be connected to the option and possibility of using the system, not necessarily to actual use. Our post hoc analyses of usage and its relation to outcomes were based on a smaller sample, comparing users and non-users, and must be viewed as an exploratory analysis only.
Interventions such as WebChoice, which offer components where people interact with each other (forum and blog), need a critical mass in order to be fully utilized. A study period of 6 months, an inclusion period of nearly 2.5 years, and the three-armed study design resulted in few participants receiving the WebChoice intervention simultaneously. This could be an explanation for low use of the discussion forum in this study.
Patients included in the study were younger (median age 52) than those who were excluded (median age 67) or who declined to participate (median age 59). In addition, the participants had higher education levels than the average level of education in Norway, suggesting that they were not representative of all age and educational groups among breast cancer patients. This, together with the small sample size, rate of declining participation, and the attrition from the study likely lowers the generalizability of the findings. On the other hand, the patients were recruited from three different hospitals across the country, which increased generalizability of findings across practice settings. Finally, nearly 50% of the participants were included right after the time of diagnosis. However, the other half was included at 1-10 months after diagnosis. As such, they were in different phases in their treatments and experienced different side effects at the time of questionnaire completion. This might have influenced the symptoms reported on the MSAS and HADS. On the other hand, this variability might strengthen our results.
This study illustrates the feasibility of offering parts of Web-based support systems in regular care, as the IPPC components were answered by nurses/physicians at the hospital where the patients were treated, providing the patients with easy access to the expertise, without a face-to-face appointment. The IPPC service, with its ability to reduce depression scores, will be an important component to integrate in Web-based support systems and can also be offered as a standalone system.
The integration of Web-based support into clinical practice will require some changes, and changing routines in care is challenging [
In the current study, the e-messages were primarily answered by nurses and passed on to physicians only if needed, which indicates that the IPPC can successfully be managed by nurses on the front line. Nurses are described as having a holistic approach to patients, focusing on emotional issues, consequences of disease, and illness information [
WebChoice, with all its features, had added effects compared with the IPPC alone. However, development and updating of systems such as WebChoice require far more resources than an IPPC service alone. Furthermore, IPPCs can be used for different patient groups independent of diagnosis, as the patient and provider are the ones who create the content.
An aspect that remains to be tested is whether Web-based support systems are more effective when health care personnel with treatment responsibilities for the patients answer messages within the system, rather than health care personnel without this knowledge (patients can send e-messages anonymously). To obtain a deeper understanding of experience with the use of IPPC in routine care, we are currently interviewing nurses and physicians who have answered e-messages in this study. This experience is important to guide the implementation processes in the future. In addition, as most studies report on services between patients and physicians, more research is needed to test similar services managed by nurses. Finally, the positive effects on patients’ outcomes, despite moderate user frequencies and almost no differences detected between users and non-users, calls for further research examining how the psychological effect of simply having access to information and support might impact outcomes.
In spite of the practice variations at three different hospitals, and moderate use of the IPPC service and WebChoice by study patients, our preliminary results suggest that offering Web-based support as a part of regular practice can be a powerful tool to help patients manage their illness. Our finding that the nurse-administered IPPC alone significantly reduced depression, a highly debilitating symptom, is particularly promising, as an IPPC can be implemented in different settings. However, the multicomponent support system, WebChoice, had additional positive effects on reducing anxiety symptoms and symptom distress.
Informed consent.
CONSORT-EHEALTH checklist V1.6.2 [
Cancer Behavioral Inventory
Internet-Based Patient Provider Communication service
Hospital Anxiety and Depression Scale
linear mixed model
Memorial Symptom Assessment Scale
randomized controlled trial
Self-Administered Comorbidity Questionnaire
The study was funded by the South-East Regional Health Authority of Norway (grant number 2009051). WebChoice was provided by the Centre for Shared Decision Making and Collaborative Care Research at Oslo University Hospital, Norway. The center delivered the intervention and managed the trial.
We would like to thank all patients, nurses, physicians, and social care workers who participated in the study.
The last author (CR) is the developer of WebChoice but has no financial or ownership rights to the application. EB, MC, ME, and CR are affiliated with the developing center but have no financial or non-financial interest to declare in relation to the study.