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.
Up to 9% of young people suffer from depression. Unfortunately, many in need of help remain untreated. The Internet offers anonymous ways to help depressed youth, especially those who are reluctant to search for help because of fear of stigma.
Our goal was to evaluate the effectiveness of an individual chat treatment based on Solution-Focused Brief Therapy (SFBT) to young individuals aged 12-22 years with depressive symptoms by comparing it to a waiting list control group.
For this study, 263 young people with depressive symptoms were randomized to the Web-based SFBT intervention, PratenOnline, or to a waiting list control condition. The chat treatment was delivered by trained professionals. Groups were compared on depressive complaints as measured by the Center for Epidemiologic Studies Depression Scale (CES-D) after 9 weeks and 4.5 months. For the chat group only, changes in depressive symptoms at 7.5 months after baseline were explored.
The experimental SFBT condition (n=131) showed significantly greater improvement than the waiting list condition (n=132) in depressive symptoms at 9 weeks and 4.5 months on the CES-D, with a small between group effect size at 9 weeks (
The Web-based SFBT chat intervention of PratenOnline was more effective than a waiting list control group in reducing depressive symptoms, and effects were larger at follow-up then at post-treatment. More studies are needed to find out if outcomes will be replicated, especially for those younger than 18 year old.
Netherlands Trial Register: NTR 1696; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=1696 (Archived by WebCite at http://www.webcitation.org/6DspeYWrJ).
Depression is among the most common mental health problems in young people. About 5.6% of youth aged 13-18 [
Despite the high prevalence of depression in youth and the possible serious implications on their lives, depression in young people is often unrecognized and undertreated [
An increasing number of Web-based services and interventions are available for children, adolescents, and young adults ranging from self-help materials to online treatments. Research on youth and young adults indicates that Web-based interventions can be effective in reducing depressive complaints [
In this paper, we present the results of a trial on a Web-based anonymous SFBT chat intervention for depressed adolescents and young adults aged 12-22 years. The trial was started after a pilot study showed promising results: a positive evaluation by participants and a decrease from pre- to post-intervention with a large effect size (
A randomized controlled trial with two parallel groups was conducted to evaluate the effectiveness of the PratenOnline chat intervention (Chat) by comparing it to a waiting list control group (WL). This study was registered with the Netherlands Trial Register (NTR 1696). Ethical approval was granted by an independent medical ethics committee (Centrale Commissie Mensgebonden Onderzoek, CCMO No. NL25219.097.08).
Participants were young people with depressive symptoms who fulfilled the following criteria: (1) 12-22 years of age, (2) had access to a computer and Internet, (3) had a CES-D score of 22 or higher (the cut-off to detect possible cases of depression among adolescents) [
Participants for the study were recruited through articles in newspapers, and banners and links placed on relevant websites for youth and on Facebook. Young people interested in participating were referred to the PratenOnline website for information about the study and to fill in a screening questionnaire to check the criteria for involvement. Those aged 12-17 years with a CES-D score of 22 or higher were invited to fill in a Web-based informed consent form and baseline questionnaire. Candidates younger than 18 years also needed written parental consent. After inclusion, participants were automatically randomized to one of two conditions: the PratenOnline chat intervention (Chat) or the waiting list control condition (WL). Random allocation was automated by a computer program without interference of the intervention supervisor or researcher. Participants were informed by email of their allocation, and the Chat participants were asked to schedule their first chat session via the intervention website. During the study, the PratenOnline chat intervention was exclusively accessible for applicants participating in the study. Blinding of participants, therapists, and researchers was not possible due to the design of this study. During the trial, participants in both conditions were allowed to seek additional help if they wished.
The intervention is a brief Web-based Solution-Focused synchronous chat intervention for young people aged 12-22 with depressive symptoms called PratenOnline (Talking online) [
The waiting list (WL) group did not receive access to the chat intervention. They could participate after the waiting period of 4.5 months.
Assessments took place before randomization (baseline, t0), 9 weeks (t1), and 4.5 months after baseline (t2). At 7.5 months after baseline (t3), a last follow-up measurement took place, exclusively for participants in the Chat condition, to measure effects at longer term. All assessments consisted of self-reported Web-based questionnaires and took about 15 minutes to complete. Email reminders were sent after 7 days if necessary. To stimulate response, participants received a voucher of €10 for each completed questionnaire (t1, t2, and t3).
Symptoms of depression in the past week were assessed with the 20-item CES-D [
At baseline, demographic characteristics (ie, sex, age, educational level, daily activity, living situation, ethnic background), duration of the psychological complaints (ie, how long the current complaints had been present), and professional help received ever before and at present were assessed. At t1, professional help and use of medication were measured. Attendance of chats was automatically measured by client Web statistics.
Originally, the trial was powered to detect clinically significant health gains expressed as a standardized effect size of a medium size (difference between groups of at least
All analyses were performed on the intention-to-treat sample with missing values imputed. The expectation-maximization (EM) method was used to impute missing data. It imputes values by maximum-likelihood estimation using the observed data in an iterative process [
Change scores based on EM imputation were used to analyze differences between groups at 9 weeks and 4.5 months (a positive score means improvement). Variables on which conditions differed significantly at baseline were regarded as relevant confounders when causing a change of 10% in the regression coefficient for condition when added to the regression model [
As attrition was rather high, sensitivity analyses were run to study the robustness of the estimates of EM imputation, using the multiple imputation Predictive Mean Matching method (PMM) in Stata (creating 100 datasets). PMM combines the standard linear regression and the nearest-neighbor imputation approaches. Predictors of outcome and missingness were taken into account to impute missing CES-D outcomes. Analyses were performed in a multiple imputation framework. Also data of completers of questionnaires were analyzed.
Magnitudes of intervention effects were estimated using Cohen’s
The proportion of participants showing reliable and clinically significant improvement [
The change from baseline to 7.5 months (t3) was explored in the Chat condition only by means of a one sample two-sided
The analyses were performed using SPSS (version 19.0) and Stata (version 11.1).
Participants were recruited from August 9, 2009, until January 24, 2010. Most participants were recruited via Internet (520/592, 87.8%). Others applied on advice of a person (61/592, 10.3%) or after reading about it in a magazine or newspaper (11/592, 1.9%).
Of the 592 young people who applied, 263 (44.4%) were included in the study. Reasons for non-inclusion were lack of informed consent (265/329, 80.5%), not completing the t0 questionnaire (39/329, 11.9%), and a CES-D depression score lower than 22 (25/329, 7.6%). Only 10 (3.8%) participants included were between 12 and 17 years of age (five assigned to each arm). Of the 253 applicants between 12 and 17 years of age, 243 were excluded, either because they did not return their parents’ consent (227/243, 93.4%) or had a CES-D score lower than 22 (16/243, 6.6%).
Flow of participants.
Baseline demographic, psychosocial, and clinical characteristics are shown in
Baseline characteristics (n=263).
Characteristics | Chat n=131 | WL n=132 | All N=263 | Statistics | |
Female, n (%) | 104 (79.4) | 103 (78.0) | 207 (78.7) | χ2
1=0.07, |
|
Age, mean (SD) | 19.4 (1.6) | 19.6 (1.8) | 19.5 (1.7) |
|
|
|
|
|
|
χ2
1=0.00, |
|
|
12-17 | 5 (3.8) | 5 (3.8) | 10 (3.8) |
|
|
18-22 | 126 (96.2) | 127(96.2) | 253 (96.2) |
|
|
|
|
|
χ2
2=2.75, |
|
|
Low | 75 (57.3) | 78 (59.1) | 153 (58.2) |
|
|
Middle | 50 (38.2) | 42 (31.8) | 92 (35.0) |
|
|
High | 6 (4.6) | 12 (9.1) | 18 (6.8) |
|
|
|
|
|
χ2
2=0.57, |
|
|
Student (high school) | 91 (69.5) | 96 (72.7) | 187 (71.1) |
|
|
Paid job | 20 (15.3) | 20 (15.2) | 40 (15.2) |
|
|
Other | 20 (15.3) | 16 (12.1) | 36 (13.7) |
|
|
|
|
|
χ2
3=0.66, |
|
|
With parents | 82 (62.6) | 78 (59.1) | 160 (60.8) |
|
|
With partner | 18 (13.7) | 17 (12.9) | 35 (13.3) |
|
|
Alone | 15 (11.5) | 18 (13.6) | 33 (12.5) |
|
|
With others | 16 (12.2) | 19 (14.4) | 35 (13.3) |
|
Had professional help before, n (%) | 68 (51.9) | 67 (50.8) | 135 (51.3) | χ2
1=0.04, |
|
Had professional help at baseline, n (%) | 26 (19.8) | 18 (13.6) | 44 (16.7) | χ2
1=1.82. |
|
Ethnic backgroundb, n (%) | 19 (14.5) | 12 (9.1) | 31 (11.8) | χ2
1=1.85, |
|
Duration psychological complaints in years, n (%) | 0.70 (1.1) | 0.52 (0.8) | 0.61 (1.0) |
|
|
CES-D depression score, mean (SD) | 39.5 (8.6) | 39.7 (7.1) | 39.6 (7.9) |
|
aEducation: lower=primary education or lower general secondary education, middle=intermediate vocational or high school, high=higher vocational education or university.
bNon-western immigrants when one or both parents is born in Africa, Latin America, or Asia (including Turkey and excluding Indonesia, Japan, and Dutch East Indies).
A total of 42.2% (111/263) of the participants did not complete t1, and 49.0% (129/263) did not complete t2. The groups did not differ at t1 in returning completed questionnaires (χ2
1=0.03,
The results for the CES-D outcomes for the intention-to-treat sample are depicted in
Means and estimates for depression score (CES-D) at 9 weeks (t1) and 4.5 months (t2) follow-up: intention-to-treat (EM imputation) and completers only (CO) analysis.
|
Chat | WL | Between group test of change scores | |||||||
|
N | Mean (SD) |
|
N | Mean (SD) |
|
Δ |
|
df |
|
T0 | 131 | 39.49 (8.58) |
|
132 | 39.74 (7.13) |
|
|
|
|
|
T1 EM | 131 | 29.20 (10.66) | 1.20 | 132 | 32.51 (9.68) | 1.01 | 0.18 | 2.70 | 261 | .007 |
T2 EM | 131 | 24.86 (8.51) | 1.72 | 132 | 33.09 (9.69) | 0.93 | 0.79 | 6.39 | 261 | <.001 |
T1 PMM | 131 | 29.78 (14.41) | 1.13 | 132 | 32.78 (12.95) | 0.98 | 0.16 | 1.78 | 151.8 | .08 |
T2 PMM | 131 | 26.36 (14.81) | 1.53 | 132 | 32.99 (13.26) | 0.95 | 0.58 | 3.45 | 134.9 | .001 |
T1 CO | 74 | 29.49 (12.11) | 1.07 | 77 | 33.00 (10.95) | 0.94 | 0.14 | 1.53 | 149 | .13 |
T2 CO | 58 | 24.66 (10.87) | 1.57 | 76 | 33.37 (11.32) | 0.82 | 0.75 | 3.66 | 132 | <.001 |
a
Means on CES-D per measurement for Chat (n=131) and WL (n=132) (EM-imputed data).
At 7.5 months (t3), the mean CES-D score of the Chat group was 20.31 (SD 10.06) showing a mean change of 19.18 points since baseline (
At 9 weeks, 22.1% (29/131) participants in the Chat condition and 13.6% (18/132) in the WL condition showed reliable and clinically significant change. This difference between conditions was not significant (χ2
1=3.24,
The number of sessions attended by the subjects in the Chat condition is shown in
According to client Web statistics, 55.7% (73/131) logged into the appointment system and 42.0% (55/131) actually had one or more chats, and 58.0% (76/131) did not have any chats. At t1 and t2, not all participants in the Chat condition had completed their therapy (n=14 had chat sessions after t1 and n=6 after t2). There were no significant differences in changes in depressive symptoms between those who attended at least 1 chat session and those who had none (at 9 weeks:
Number of chat sessions attended by percentages of participants (n=131).
The present study shows considerable improvements in depressive symptoms in both the Chat group and the waiting list group over time, but more so for the SFBT chat group, indicating it was more effective than the waiting list control condition. Between group effect sizes were small at 9 weeks (
There are no Web-based studies on SFBT interventions to compare our results with, but effect sizes reported for offline SFBT in the meta-analysis of Kim (2008) [
When compared to outcomes of face-to-face treatments for adolescents found in a meta-analytic review [
In our study, attrition (ie, dropping out of the study) was high. This is a phenomenon often observed in studies of Web-based interventions both among adults and youngsters [
In our study, only 42% of those who had access to the Chat intervention of PratenOnline made use of it, although 56% had made an appointment for a chat. Limited adherence is not uncommon in Web-based interventions [
In the daily chat practice of PratenOnline, the age group of 12-17 years old is highly represented. In the trial, however, only 10 (3.8%) participants in that age group were included. The major bottleneck to participate for this age group was the parental consent that had to be provided by both parents in a written consent form. Other studies also had problems with recruitment because of parental consent [
In our study, a waiting list control condition was used. This might have effected the magnitude of the between group effect sizes. As Clarke et al pointed out, “the between group effect size is not just a function of the potency of the experimental intervention but is also a function of the magnitude of change observed in the control condition” (2009, page 231) [
As far as we know, this is the first study on a brief Web-based Solution-Focused Intervention for young people with depressive complaints. Despite the limitations of the present study, our findings indicate that adolescents and young adults with depressive symptoms can profit from access to the Web-based SFBT chat treatment. Studies in this field are few, but this one contributes to the evidence that Web-based interventions can be effective. However, because of the limitations of the study, more research is needed to find out if outcomes will be replicated. Especially for young people under the age of 18, more evidence is needed for the effectiveness of Web-based SFBT. To make such studies successful, the major impediment to include this age group needs to be tackled: the parental consent.
CONSORT-EHEALTH checklist V1.6.2 [
Center for Epidemiologic Studies Depression Scale
completers only
expectation-maximization
predictive mean matching
Quick Inventory of Depressive Symptomatology—Self rated
Solution-Focused Brief Therapy
waiting list
The study was funded by a research grant from ZonMw (Netherlands Organization for Health Research and Development), Grant No. 15700.2003. No funding bodies had any role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Also, we would like to thank all participants, without whom our study would not have been possible.
Pien Oijevaar was one of the founders of the chat therapy, PratenOnline, in the Netherlands, but she did not derive financial income from the PratenOnline intervention. All other authors declare no conflicts of interest.