The study is a cluster-randomized, controlled, non-blinded, multicenter study with two parallel groups to be compared (Fig. 1). The study investigates the effects of two different ways of assessing psychosocial distress in outpatient HGG patients.
The clusters are the institutions (hospitals) participating in this trial (n=13). In intervention arm hospitals, distress is assessed within the doctor-patient consultations. In CAU hospitals distress is assessed with the DT.
Patients with HGG will be recruited from thirteen geographically dispersed neuro-oncological centers throughout Germany (Tübingen, Mainz, Frankfurt, Ludwigsburg, Würzburg, München, Stuttgart, Köln, Düsseldorf, Berlin, Ulm, Trier, Leipzig). The time course is displayed in figure 2.
The allocation to the two study arms is randomized. Through cluster randomization, the study is easier to implement into clinical routine compared to individual patient randomization. An external statistician who is not involved in treatment, planning or evaluation of the study has performed the randomization (Center for Clinical Studies in Mainz, Germany).
Participants and procedures
a) Clusters: Departments as clusters were included, when 1) a neuro-oncological outpatient service was present with 2) regular weekly consultation hours and 3) conducted by neurologists or neurosurgeons.
b) Patients: We include patients with a HGG (glioblastoma WHO°IV, anaplastic astrocytoma WHO°III, anaplastic oligodendroglioma WHO°III) aged 18 years and older with the ability to give informed consent. Patient exclusion criteria are absence of written informed consent and inability to understand the German language.
Screening and informed consent procedure
The study coordinator or research assistant of every study center will screen the patients scheduled in the outpatient department to assess their eligibility. Then the study nurse will give each eligible patient a brief overview of the study and ask whether or not he or she would be interested in participation. The consent to participate in the study goes along with signing the consent form. In addition to the patient information, each patient receives a copy of the consent form. From decliners of the study, gender, tumor diagnosis, age and the reason for decline will be documented (Fig. 3).
Clinical information assessed in the study
The following information is documented in both arms: location of the tumor, clinical condition defined by Eastern Co-operative of Oncology Group (ECOG)- status and Karnofsky Performance Score (KPS), stage of disease (initial diagnosis/recurrence), neurological deficit defined by Neurologic Assessment in Neuro-Oncology (NANO) scale(16), time since initial diagnosis, age and gender. Furthermore, patients will provide demographic information such as professional and financial situation, family status and current care situation.
Conditions of the intervention arm
Psychosocial distress is measured by three questions that emerged as important in a preliminary study. The three questions are:
- Has your mood worsened due to the disease?
- Do physical changes due to the disease, such as numbness, weakness or feeling exhausted more quickly, put a strain on you?
- Has your mental capacity worsened as a result of the disease, making it harder for you to concentrate or remember things, for example?
On the basis of the patient's answers, the doctor offers psychosocial care, if indicated and desired by the patient and refers the patient to the appropriate specialist services, e.g. psycho-oncological service, clinic social service, outpatient cancer counselling centers, physiotherapy, established psychotherapists, etc.
Conditions of the control arm
In the control arm, psychosocial distress is assessed using the Distress Thermometer (DT), which can be evaluated within a few minutes, so that the results in the control group can also be discussed during the doctor-patient consultation. If there is an obvious need for treatment or if a patient requires psychosocial support, the referral is carried out analogously to the intervention condition.
Psychosocial care in both arms
Depending on the urgency, care must be provided quickly in both arms, i.e. within a few days to a maximum of 2 weeks, depending on the capacity of the respective cooperating services (psycho-oncological, social-legal, etc.).
Instruments applied in the study
Before the doctor-patient-consultation, t1
First patients complete a baseline questionnaire (t1) assessing use of and wish for psychosocial support using an adapted version of a self-developed questionnaire (17). Further, the European Organization of Research and Treatment of Cancer Core Questionnaire (EORTC QLQ-C30) with its brain module (BN20) will be applied. It is a self-assessment tool to evaluate cancer patients’ quality of life. Validity and reliability have been previously tested in several clinical studies (18, 19).
Patients in the control group complete additionally the Distress-Thermometer (DT). This is a self-reporting screening instrument, measuring psychosocial distress by a numerical rating scale accompanied by a 40-item list with problems from different areas of life, e.g. practical, family, emotional problems, etc. (5, 20).
Afterwards, the patients have the doctor-patient consultation. In the intervention group, the three questions, described above, are asked during doctor-patient-consultation instead of using the DT, see also table 1.
After the doctor-patient-consultation, t2
After the doctor-patient-consultation, the patients complete the above mentioned questionnaire for support needs a second time.
In addition, the Structured Clinical Interview (SCID) for DSM (diagnostic and statistical manual of mental disorders) is conducted with the patients. The SCID is used to identify mental health conditions (21). The interviewer asks certain questions and codes the answers in order to be able to make a diagnosis at the end. The SCID is widespread, reliable and valid (21). The study nurses have been trained beforehand to conduct the interviews. Each controversial case will be discussed in a telephone conference with a supervisor, see also table 1.
Three month later, the patients will complete the EORTC QLQ-C30 + BN20 questionnaire and again an adapted version of the self-developed questionnaire (see also table 1) in order to assess the proportion of patients receiving adequate care.
Training of study personnel and monitoring
During the start-up phase of the study, the research assistants and study nurses have been trained to conduct the SCID-interview reliably. They also learned about the assessment procedures at t1-t3. Furthermore, worksheets and case report forms (CRFs) have been developed, evaluated and adapted after a pilot test in September 2019. All centers will be initiated and monitored by the project manager of the study. Frequently (monthly) conducted telephone conferences regarding the SCID-interview and study meetings planned twice a year will help to solve any problems.
General considerations for data analysis and statistical methods of the study
Through cluster randomization, the study can be better implemented into clinical routine.
Broadly selected inclusion criteria ensure that as many patients as possible are reached. The aim is to include not only highly selected patients.
Data will be derived from patients’ responses to questionnaires, worksheets completed by study nurses and research assistants as well as from medical records. All data will be collected specifically for research purposes.
All person-identifying data are avoided in the data set and code numbers used instead. A study code will be assigned to each subject in the respective study center. A copy of all worksheets and completed questionnaires using the code numbers will be sent to the statistical center (IMBEI) in Mainz. Only pseudonymized data will be handed over. All study data will be stored for 10 years after completion of the study in the respective center, and only the study team will have access to the data. In Mainz the central data processing and evaluation will be done. Pseudonymized data will be stored in a password-protected electronic database, which is stored in a locked server only accessible for staff members of the IMBEI with access authorization.
Based on previous studies (22) we expect that in the control group only about 15% of mentally stressed patients receive psychosocial care compared to expected 25% in the intervention group.
Due to the fact that this is a cluster-randomized study, more patients have to be included in the study.
The sample size calculation is based on recommendations by Hayes and Bennett, who developed several scenarios for cluster-randomized studies and provided formulas for the calculation of the cluster effect. The intracluster correlation coefficient (ICC) in a comparable, cluster-randomized study with 13 clusters was between 0.03 and 0.19 depending on the outcome(22). However, Donner and Klar (2004) warn of overestimating the stability and size of an ICC. Therefore, we assume an ICC of 0.005. We further assume that 50 patients will be included in the study in each clinic and that the clusters will be of the same size. In this scenario (ICC 0.005), 12 clinics with a total of n=616 patients are required, to demonstrate the expected intervention effect (25% vs. 15%) with α of 0.05 and a power of 80%.
Control of bias and confounding
Potential confounders (hospital organization, influence of individual doctors, local conditions affecting care and resources) are controlled by randomization. However, since cluster randomization may result in imbalances of baseline differences, these variables are also documented and analyzed.
Through the use of standardized questionnaires, the possibility of information bias is reduced.
It should be considered that support needs could be awakened by participation in this study. For this reason, questions are asked before and after the doctor's conversation in order to be able to identify possible changes. However, since screening for psychosocial stress is required in all cancer patients according to the German guideline on psycho-oncological diagnostics and treatment, we do not consider it ethically justifiable to establish a third study arm that does not ask about the burden and the care desired.
The analysis of the primary outcome "proportion of adequately care for patients with increased psychosocial stress" will be based on hierarchical models for cluster-randomized studies (with random intercept for the clinics).
Due to the limited number of clinics, the following variables will also be adjusted: Sex, age, ECOG stage, initial or recurrent disease, neurological deficit, time since initial diagnosis, living alone vs. in partnership. The analyses are performed according to intent to treat. The aim is to reduce a possible bias by selective drop-out. The secondary endpoints are also calculated using mixed models.
The study is performed in accordance with national law, institutional ethical standards, and the Declaration of Helsinki after approval of the study protocol by the local ethic committees (No. of the first approval: 837.179.17 (11013)). All centers participating in the study obtained an ethical approval before recruiting patients. All patients provide written informed consent prior to data assessment.
The study has been registered in the German Clinical Trials Register (No. DRKS00018079).