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 Formative Research, is properly cited. The complete bibliographic information, a link to the original publication on https://formative.jmir.org, as well as this copyright and license information must be included.
Increasingly, social media is being recognized as a potential resource for patient-generated health data, for example, for pharmacovigilance. Although the representativeness of the web-based patient population is often noted as a concern, studies in this field are limited.
This study aimed to investigate the sample bias of patient-centered social media in Dutch patients with gastrointestinal stromal tumor (GIST).
A population-based survey was conducted in the Netherlands among 328 patients with GIST diagnosed 2-13 years ago to investigate their digital communication use with fellow patients. A logistic regression analysis was used to analyze clinical and demographic differences between forum users and nonusers.
Overall, 17.9% (59/328) of survey respondents reported having contact with fellow patients via social media. Moreover, 78% (46/59) of forum users made use of GIST patient forums. We found no statistically significant differences for age, sex, socioeconomic status, and time since diagnosis between forum users (n=46) and nonusers (n=273). Patient forum users did differ significantly in (self-reported) treatment phase from nonusers (
Forum users showed no particular bias on the most important demographic variables of age, sex, socioeconomic status, and time since diagnosis. This may reflect the narrowing digital divide. Overrepresentation and underrepresentation of patients with GIST in different treatment phases on social media should be taken into account when sourcing patient forums for patient-generated health data. A further investigation of the sample bias in other web-based patient populations is warranted.
Web-based patient forums provide patients with both emotional and informational support [
Postmarket surveillance is necessary as clinical trials are of limited duration and suffer from sample bias; they often exclude older patients, patients with comorbidities, and pregnant women [
Reliance on spontaneous reports alone results in a severe underreporting of adverse drug responses (ADRs) [
However, researchers as well as patients have expressed concern about sample bias on social media [
On the basis of studies of the general population of social media users [
Although there is a large literature base on patient communication forums and the extraction of adverse drug effects, to date, the work on sample bias in web-based patient-to-patient communication is limited to 2 studies. Prior work on American patients with breast cancer [
Other studies addressed another bias that is relevant when mining social media for patient-generated health data: so-called activity bias [
As Baeza-Yates [
In this study, we investigated (1) what proportion of patients have contact with fellow patients on social media, (2) why patients abstain from engaging with web-based patient communities, and (3) to what extent there are significant demographic and clinical differences between those that use social media to converse with patients and those that do not. This study did not assess general social media use but focused specifically on the web-based communication with other patients. We defined social media as a web-based communication channel where information and messages are exchanged. When referring to
On the basis of general social media, we hypothesized that forum users will differ in demographic factors including age, sex, and socioeconomic status from nonusers. We also hypothesized that forum users will differ in marital status and have a lower level of social functioning than nonusers, in line with the social compensation model [
A cross-sectional study was conducted among Dutch patients with GIST aged ≥18 years at diagnosis, diagnosed between January 1, 2008, and December 31, 2018, in 5 GIST reference centers. Patients were selected from the Netherlands Cancer Registry (NCR), a population-based registry, which is maintained by the Netherlands Comprehensive Cancer Organization (in Dutch: Integraal Kankercentrum Nederland or IKNL) and collects patient and tumor characteristics on all newly diagnosed patients with cancer in the Netherlands. Exclusion criteria were cognitive impairment or being too ill at the time of the study according to advice from a (former) treating specialist. Eligible patients were invited by their (former) treating physician by a letter explaining the study. Upon consent of the patient, including permission to link the survey data with NCR data, patients could complete the survey on the web or on paper upon request. Refer to
CONSORT (Consolidated Standards of Reporting Trials) flow diagram of response rate.
Ethics approval for the cross-sectional study was provided by the medical ethical committee of the Radboud University Medical Centre (2019-5888). According to the Dutch law, approval of one ethical committee for questionnaire research is valid for all participating centers. Patients gave informed consent, including permission to link the survey data with NCR data, before completing the survey.
Participants completed questions regarding their participation in social media and web-based patient communities. These questions were developed by the authors. Respondents were asked whether and how patients use digital platforms to have contact with other patients. Possible answers (translated to English) were “Generic social media (like Facebook or Twitter),” “General forum or discussion group,” “Specific online patient forum,” “Other, namely...,” or “I do not use digital communication.” Patients were provided with the following definition for a digital medium (translated to English): a web-based communication channel where information and messages are exchanged between participants. Patients were allowed to give multiple answers.
Respondents having contact with other patients on the web were subsequently asked about their motivations for going on the web and about their frequency of posting messages. Both questions were adapted from a Dutch survey designed by van Uden-Kraan et al [
Demographic variables (ie, age, sex, and socioeconomic status) as well as clinical variables (ie, tumor type, tumor stage, time since diagnosis, whether surgery was performed, and whether targeted therapy was part of treatment) of survey respondents were collected from the NCR. Survey respondents were additionally asked about their marital status, their current treatment phase, whether they presently use medication, their most recent medication (if any), and the presence of the 14 comorbid conditions measured in the Charlson comorbidity index [
The options patients can choose for self-reported treatment phase are defined as follows: “Cured and not monitored” (“I am cured and no longer need to be monitored”) refers to patients who are considered cured after surgery with or without adjuvant imatinib, “On curative treatment” (“I am being treated and can still be cured”) refers to patients who are undergoing adjuvant imatinib treatment, “Follow-up after treatment” (“I am not being treated but am only being monitored”) refers to patients who are being monitored after surgery with or without adjuvant imatinib and are not undergoing treatment at this time, “On palliative treatment” (“I am being treated but cannot be cured”) refers to patients undergoing palliative treatment with thyroid kinase inhibitors, and “Best supportive care” (“I cannot be cured but am not being treated”) refers to patients who are palliative but are not receiving TKIs.
To measure overall health-related quality of life (QoL), social functioning, and symptom burden, participants completed the European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire C30 version 3.0 [
Any questions that were not previously validated were pretested with patients and changed according to their feedback (cognitive debriefing). The questionnaires cannot be shared because of copyright restrictions.
The reasons for abstaining and engaging with web-based patient-to-patient communication were analyzed manually by the first author. In total, 15.8% (52/328) of the cases contain missing data. As none of these cases are forum users, the data are not missing completely at random. As we do not observe any other patterns in the missing data that cannot be explained by the variables on which we have full information, the data are missing at random. As the missing data occur in multiple variables, we used Multivariate Imputation by Chained Equations [
We aimed to analyze whether there were statistically significant differences in demographic and clinical characteristics as well as the QoL measures between forum users and nonusers. For each imputed data set, a multiple logistic regression analysis was performed with forum use as the dependent variable and demographic and clinical factors as the independent variables (refer to Surveys). The effects of one variable on forum use are thus conditional on the other variables in the model. We report the average and SD of the 20 imputed data sets, as this provides a more reliable result than a single run. We use the mean as the average for all variables except the
For this analysis, the number of variables was restricted by the small size of the user population. We checked for multicollinearity using Variance Inflation Factor tests. If the Variance Inflation Factor value was >3, we removed one of the collinear explanatory variables. In total, we removed 2 variables accordingly: the most recent medication and whether the patient is on systemic treatment currently (“On systemic treatment currently”). Note that whether the patient received targeted therapy at some point in time (“Targeted therapy”) is included. Moreover, 2 categories of self-reported treatment phase, namely, on palliative treatment and on best supportive care, needed to be merged into one palliative category, as only one patient was receiving best supportive care.
Benjamini-Hochberg correction [
In total, 328 patients with GIST responded to the survey (response rate 64%). The median age of the participants was 67 (range 28-91) years, and 53.8% (174/328) of the participants were male (
Overall, 9 patients did not answer the question about forum use, and their forum use is thus unknown. Consequently, the sum of the reported numbers under forum use (Yes and No) does not equal the number reported for all respondents. The percentages were calculated based on the counts per category, that is, 54.9% (150/273) of nonusers are male.
Demographic characteristics of survey respondents.
Demographic characteristic | All (N=328) | Forum usera | ||
|
|
No (n=273) | Yes (n=46) | |
Age (years), median (range) | 67 (28-91) | 68 (28-91) | 65 (47-83) | |
|
||||
|
Male | 174 (53) | 150 (54.9) | 21 (45.7) |
|
Female | 154 (47) | 123 (45.0) | 25 (54.3) |
|
||||
|
Low (1-3) | 90 (27.4) | 74 (27.1) | 13 (28.3) |
|
Intermediate (4-7) | 132 (40.2) | 113 (41.4) | 16 (34.8) |
|
High (8-10) | 106 (32.3) | 86 (31.5) | 17 (37) |
|
||||
|
Married or living together | 246 (7) | 202 (74) | 38 (82.6) |
|
Single | 79 (24.1) | 68 (24.9) | 8 (17.4) |
|
Missing | 4 (1.2) | 3 (1.1) | 0 (0) |
Time since diagnosis (years), median (range) | 5 (1-12) | 5 (1-12) | 5 (2-11) | |
|
||||
|
I | 121 (36.9) | 109 (39.9) | 8 (17.4) |
|
II | 61 (18.6) | 51 (18.7) | 10 (21.7) |
|
III | 66 (20.1) | 53 (19.4) | 10 (21.7) |
|
IV | 55 (16.8) | 38 (13.9) | 16 (34.8) |
|
Missing | 25 (7.6) | 22 (8.1) | 2 (4.3) |
|
||||
|
Yes | 287 (87.5) | 244 (89.4) | 36 (78.3) |
|
No | 41 (12.5) | 29 (10.6) | 10 (21.7) |
|
||||
|
Yes | 214 (65.2) | 170 (62.3) | 39 (84.8) |
|
No | 114 (34.8) | 103 (37.7) | 7 (15.2) |
|
||||
|
Cured and not monitored | 61 (18.6) | 56 (20.5) | 2 (4.3) |
|
On curative treatment | 34 (10.4) | 31 (11.4) | 3 (6.5) |
|
Follow-up after treatment | 162 (49.4) | 139 (50.9) | 19 (41.3) |
|
On palliative treatment | 66 (20.1) | 42 (15.4) | 22 (47.8) |
|
Best supportive care | 1 (0.3) | 1 (0.4) | 0 (0) |
|
Missing | 4 (1.2) | 4 (1.5) | 0 (0) |
|
||||
|
Yes | 208 (63.4)b | 181 (66.3) | 25 (54.3) |
|
No | 108 (32.9) | 83 (30.4) | 21 (45.7) |
|
Missing | 12 (3.7) | 9 (3.3) | 0 (0) |
|
||||
|
Imatinib | 178 (54.3) | 140 (51.3) | 31 (67.4) |
|
Sunitinib | 9 (2.7) | 7 (2.6) | 2 (4.3) |
|
Regorafenib | 6 (1.8) | 4 (1.5) | 2 (4.3) |
|
Other | 15 (4.6) | 8 (2.9) | 4 (8.7) |
|
No therapy | 114 (34.8) | 103 (37.7) | 7 (15.2) |
|
Missing | 14 (4.3) | 11 (4) | 0 (0) |
|
||||
|
0 | 109 (33.2) | 92 (33.7) | 14 (30.4) |
|
1 | 71 (21.6) | 59 (21.6) | 10 (21.7) |
|
2+ | 146 (44.5) | 120 (44.0) | 22 (47.8) |
|
Missing | 2 (0.6) | 2 (0.7) | 0 (0) |
Global health scale (0-100), mean (SD) | 78.6 (18.1) | 79.0 (17.7) | 76.1 (20.1) | |
Symptom burden (0-100), mean (SD) | 12.1 (12.8) | 11.4 (12.6) | 15.6 (13) | |
Social functioning (0-100), mean (SD) | 92.4 (18.9) | 93.8 (17.1) | 84.8 (26) |
aNine participants did not answer this question.
bIt appears that patients who are currently being monitored may have misunderstood this question, inflating the number of patients who are currently on targeted medication for gastrointestinal stromal tumor.
Among the participating Dutch patients with GIST, 81% do not have contact with other patients via any social media platform (
Descriptive statistics for the use of social media to have contact with other patients (N=328).
Which of the following digital media do you use to have contact with other patients?a (Indicate all that apply) | Values, n (%) |
General social media (such as Facebook or Twitter) | 6 (1.8) |
General cancer-related forum or discussion group | 7 (2.1) |
GISTb-specific web-based patient forum | 46 (14) |
Any social medium | 59 (18) |
None or via another medium than social media | 265 (80.8) |
Missing | 4 (1.2) |
aRespondents can give multiple answers to this question.
bGIST: gastrointestinal stromal tumor.
The reasons nonusers report for not using social media to communicate with other patients (N=265).
Self-reported reasona | Values, n (%) |
Feel no need to communicate (digitally) with other patients | 78 (29.4) |
I find it too confronting or burdensome | 33 (12.5) |
I do not know where to find online communities | 30 (11.3) |
There are too many negative comments | 26 (9.8) |
I do not have the time | 23 (8.7) |
The information shared is useless or less valuable | 20 (7.5) |
I communicate with enough patients personally or via another nondigital medium | 18 (6.8) |
I do not use social media, lack a computer or digital skills, or do not like obtaining information digitally | 8 (3) |
I obtain sufficient information via my medical specialist or by searching online | 7 (2.6) |
I no longer have symptoms or do not like to consider myself a patient | 5 (1.9) |
I have privacy concerns | 3 (1.1) |
They do not exist in my language | 2 (0.8) |
No particular reason | 1 (0.4) |
Missing | 20 (7.5) |
Survey respondents most frequently used patient forums to communicate with other patients. The number of survey responders that made use of other web-based platforms was too small to analyze how they compare with nonusers. Thus, we will focus on analyzing the sample bias of GIST-specific patient forums. Hereafter, when we refer to “forum users,” we mean users of GIST-specific patient forums.
The reasons users report for visiting the patient forum (N=45).
Self-reported reasona | Value, n (%) |
When I have a question about my illness | 18 (40) |
When I have heard new information about my illness | 18 (40) |
When I am curious about how other members are doing | 16 (36) |
When I observe new symptoms | 14 (31) |
When I have a lot of symptoms | 6 (13) |
When I feel insecure | 5 (11) |
Before making a medical choice | 4 (9) |
Because I enjoy the company | 4 (9) |
Because other members expect me to be there | 2 (4) |
When I feel lonely | 1 (2) |
It is part of my daily routine | 1 (2) |
I never use the forum anymore | 1 (2) |
In total, 85.8% (273/328) of the participants were not making use of specialized GIST patient forums (
We did not find significant differences between forum users and nonusers for other disease-related characteristics when they were adjusted for covariates. We also did not find significant differences in key demographic variables such as age, sex, socioeconomic status, and marital status. However, we did find a significant difference in level of social functioning in 7 of 20 runs (LR=6.8;
Average results (with SD) of a logistic regression of demographic and clinical characteristics of patient forum users and nonusers using Multivariate Imputation by Chained Equations with 20 runs.
|
Coefficient, mean (SD) | SE, mean (SD) |
|
LR, mean (SD) | Odds ratio | ||||||||||||
|
|
|
|
|
|
5% (SD) | Mean (SD) | 95% (SD) | |||||||||
Intercept | −2.80 (0.54) | 2.08 (0.03) | N/Ab | N/A | N/A | N/A | N/A | N/A | |||||||||
Age | −0.02 (0.004) | 0.02 (0.0002) | 1 | 1.32 (0.54) | .26 (.10) | 0.95 (0.004) | 0.98 (0.004) | 1.02 (0.004) | |||||||||
Sex | 0.62 (0.04) | 0.37 (0.004) | 1 | 2.86 (0.35) | .09 (.02) | 0.90 (0.03) | 1.86 (0.07) | 3.86 (0.16) | |||||||||
|
N/A | N/A | 2 | 1.37 (0.49) | .25 (.08) | N/A | N/A | N/A | |||||||||
|
Low (1-3)c | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | ||||||||
|
Intermediate (4-7) | −0.39 (0.10) | 0.44 (0.006) | N/A | N/A | N/A | 0.20 (0.03) | 0.68 (0.07) | 1.62 (0.16) | ||||||||
|
High (8-10) | 0.05 (0.10) | 0.44 (0.005) | N/A | N/A | N/A | 0.45 (0.04) | 1.06 (0.10) | 2.50 (0.26) | ||||||||
Marital status | −0.32 (0.09) | 0.47 (0.006) | 1 | 0.52 (0.25) | .47 (.11) | 0.29 (0.04) | 0.73 (0.06) | 1.82 (0.06) | |||||||||
Time since diagnosis | 0.02 (0.02) | 0.07 (0.001) | 1 | 0.12 (0.16) | .85 (.15) | 0.88 (0.02) | 1.02 (0.02) | 1.17 (0.02) | |||||||||
Tumor type | 0.57 (0.06) | 0.38 (0.003) | 1 | 2.29 (0.52) | .13 (.04) | 0.84 (0.05) | 1.77 (0.11) | 3.70 (0.24) | |||||||||
|
N/A | N/A | 3 | 2.60 (0.92) | .12 (.07) | N/A | N/A | N/A | |||||||||
|
Ic | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | ||||||||
|
II | 0.51 (0.13) | 0.55 (0.009) | N/A | N/A | N/A | 0.57 (0.07) | 1.67 (0.21) | 4.89 (0.63) | ||||||||
|
III | 0.21 (0.21) | 0.63 (0.01) | N/A | N/A | N/A | 0.37 (0.09) | 1.27 (0.29) | 4.31 (0.94) | ||||||||
|
IV | 0.86 (0.17) | 0.66 (0.01) | N/A | N/A | N/A | 0.66 (0.12) | 2.41 (0.43) | 8.83 (1.61) | ||||||||
Surgery | 0.04 (0.12) | 0.57 (0.01) | 1 | 0.05 (0.10) | .89 (.10) | 0.34 (0.04) | 1.05 (0.12) | 3.24 (0.42) | |||||||||
Targeted therapy | 0.12 (0.10) | 0.57 (0.01) | 1 | 0.07 (0.08) | .83 (.10) | 0.37 (0.03) | 1.13 (0.11) | 3.49 (0.38) | |||||||||
|
N/A | N/A | 3 | 10.67 (1.10) | N/A | N/A | N/A | ||||||||||
|
Cured and not monitoredc | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | ||||||||
|
On curative treatment | 0.59 (0.26) | 1.07 (0.05) | N/A | N/A | N/A | 0.23 (0.04) | 1.86 (0.45) | 15.56 (4.65) | ||||||||
|
Follow-up after treatment | 1.03 (0.26) | 0.87 (0.06) | N/A | N/A | N/A | 0.52 (0.08) | 2.88 (0.69) | 16.18 (5.10) | ||||||||
|
Palliative | 2.29 (0.23) | 0.97 (0.06) | N/A | N/A | N/A | 1.50 (0.23) | 10.11 (2.21) | 68.68 (19.84) | ||||||||
|
N/A | N/A | 2 | 0.42 (0.26) | .53 (.14) | N/A | N/A | N/A | |||||||||
|
0c | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | ||||||||
|
1 | 0.28 (0.11) | 0.50 (0.007) | N/A | N/A | N/A | 0.50 (0.06) | 1.33 (0.14) | 3.51 (0.36) | ||||||||
|
2+ | 0.21 (0.08) | 0.45 (0.005) | N/A | N/A | N/A | 0.51 (0.04) | 1.23 (0.09) | 2.99 (0.24) | ||||||||
Global health scale or QoLd | 0.03 (0.002) | 0.01 (0.0001) | 1 | 4.38 (0.69) | .04 (.02) | 1.00 (0.002) | 1.04 (0.002) | 1.06 (0.002) | |||||||||
Symptom burden | −0.0003 (0.005) | 0.02 (0.0004) | 1 | 0.09 (0.10) | .83 (.11) | 0.96 (0.006) | 1.00 (0.005) | 1.04 (0.005) | |||||||||
Social functioning | −0.03 (0.002) | 0.01 (0.0002) | 1 | 6.87 (0.90) | 0.96 (0.002) | 0.98 (0.001) | 0.99 (0.002) |
aThere are separate
bN/A: not applicable.
cThese categories were taken as the reference categories for calculating the influence of different categories of the variable on forum use
dQoL: quality of life.
*Italicized values indicate statistically significant values.
A survey was conducted among 328 patients with GIST in the Netherlands. Our results show that most survey respondents do not have contact with other patients via social media. They indicate a large heterogeneity of reasons of why they abstain from doing so, with the most prevalent being they feel no need, find it too confronting, or do not know where to find such web-based communities. Of the minority who do use social media for this purpose, most use disease-specific patient forums. The most prevalent reasons for accessing a patient forum are (1) having a question about their illness, (2) having heard new information, (3) experiencing new symptoms, or (4) wondering how other patients are doing. Patient forum users differ significantly in (self-reported) treatment phase from nonusers. Patients in the palliative phase are 10 times more likely to be forum users than patients who are cured. Patients who are monitored approximately 3 times and patients undergoing curative treatment approximately 2 times are more likely to be users than cured patients. For 7 of 20 data imputations, forum users also have a significantly lower level of social functioning.
In contrast to the general population of social media users, patient forum users do not appear to differ in age, sex, and socioeconomic status from nonusers. On the one hand, this may be an effect of the increasingly more widespread adoption of social media. This idea is supported by the small number of patients that indicate they lack the skills or access to be on social media (8/265, 3.3%). On the other hand, it is also possible that there is less demographic bias on patient forums than in general social media. This may be related to the widely different goals that users have with their participation. Although a feeling of community and social support may overlap, patients report motivations such as questions around their illness and the experience of new symptoms that normal social media users are unlikely to share.
Prior work [
First and foremost, we only studied a specific patient population in a single country, and thus, further research is needed to elucidate to what extent our results are generalizable. Patients in other countries may have lower digital access or skills or may not wish to use social media for patient-to-patient communication for other reasons (eg, other privacy laws or country-specific customs).
Our choice of patients with GIST as a target population may also impact to which disorders our results generalize to. Patients with GIST have a median age of mid-60 years [
A second limitation of this study is the small sample size. Among the 328 respondents, only 46 (14.1%) indicate that they make use of patient forums. Nonetheless, given the low incidence of GIST at 12.7 per million [
On the basis of this work, a number of recommendations can be made. First, of the possible digital resources that can be used to source complementary real-world evidence, for instance, for pharmacovigilance, patient forums should be preferred over other social media. Our results reveal that patients with GIST strongly prefer disease-specific patient forums over general social media for communicating with fellow patients. However, most research in this field currently focuses on general social media such as Twitter [
Although we find that there is sample bias in patient forum users and, thus, the sample is not wholly representative for the patient population, sample bias is also a concern for other sources of patient reports. Understanding which patients are overrepresented and underrepresented on web-based forums is the first step to using web-based patient reports as a complementary resource, for instance, for pharmacovigilance, which is seen as a realistic first use case. For pharmacovigilance specifically, it is not of great concern that patients who are considered cured and not undergoing treatment currently are underrepresented. Future work into comparing the sample bias of clinical trials with that of web-based patient forums would be beneficial to further explore its complementary value in detail. It would also be valuable to gain more insight into the different types of forum users.
Second, it may be beneficial to create awareness among medical professionals that patients are more likely to search for information in web-based patient communities when they have questions, have been given new information, or have new symptoms. Medical professionals could try to aid patients in their information need by pointing them toward such resources in these cases. This may also take away the barrier mentioned by patients that they do not know where to find such web-based communities.
Third, future work into the sample bias of patient forums for other patient populations is necessary, as this study was limited to a single population in a single country. Nonetheless, our work is a stepping stone toward dissuading the concerns that researchers have expressed regarding the sample bias of social media [
In this study, we investigated how representative participants in patient forums are for the general patient population by conducting a survey among patients with GIST in the Netherlands. We found statistically significant differences in terms of treatment phase and offline social support between forum users and nonusers. The consequent overrepresentation and underrepresentation of certain types of patients should be considered when sourcing patient forums for patient-generated health data. As our study was limited to a single patient population, a further investigation of the sample and activity bias in other web-based patient populations is warranted. Sample bias is inherent to any information source, and only through awareness of these biases can these resources be used as a source for complementary real-world evidence in the future.
adverse drug response
gastrointestinal stromal tumor
likelihood ratio
Netherlands Cancer Registry
quality of life
tyrosine kinase inhibitor
This survey was partly funded by a research grant from Novartis (grant 006.18). AD is supported by a research grant from the Stichting Internet Domeinregistratie Nederland fund (grant 174054). DdH is supported by a research grant from the European Organisation for Research and Treatment of Cancer Quality of Life Group (grant 003-2018). The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. The authors would like to thank Carla Fokkema-Vlooswijk and Esther Derksen-Peters for their support with the recruitment and data collection through Patient-Reported Outcomes Following Initial Treatment and Long-term Evaluation of Survivorship in the survey study.
The data are not publicly available.
AD, DdH, SV, and OH contributed to conceptualization. AD, DdH, SV, OH, WTAvdG, and GvO contributed to methodology. AD, DdH, ID, OH, WTAvdG, AO, AKLR, NS, WvL, and HG contributed to formal analysis and investigation. AD wrote the original paper. DdH, SV, ID, OH, WTAvdG, AO, AKLR, NS, WvL, GvO, HG, and WK reviewed and edited the article.
None declared.