Article Text
PDF
Abstract
Background Patient-reported experience measures (PREMs) are valuable tools to evaluate patient-centredness (PC) from the patients’ perspective. Despite their utility, a comprehensive PREM addressing PC has been lacking. To bridge this gap, we developed the preliminary version of the Experienced Patient-Centeredness Questionnaire (EPAT), a disease-generic tool based on the integrative model of PC comprising 16 dimensions. It demonstrated content validity. This study aimed to test its psychometric properties and to develop a final 64-item version (EPAT-64).
Methods In this cross-sectional study, we included adult patients treated for cardiovascular diseases, cancer, musculoskeletal diseases and mental disorders in inpatient or outpatient settings in Germany. For each dimension of PC, we selected four items based on item characteristics such as item difficulty and corrected item–total correlation. We tested structural validity using confirmatory factor analysis, examined reliability by McDonald’s Omega and tested construct validity by examining correlations with general health status and satisfaction with care.
Results Analysis of data from 2.024 patients showed excellent acceptance and acceptable item–total correlations for all EPAT-64 items, with few items demonstrating ceiling effects. The confirmatory factor analysis indicated the best fit for a bifactor model, where each item loaded on both a general factor and a dimension-specific factor. Omega showed high reliability for the general factor, while varying for specific dimensions. Construct validity was confirmed by absence of strong correlations with general health status and a strong correlation of the general factor with satisfaction with care.
Conclusions EPAT-64 demonstrated commendable psychometric properties. This tool allows comprehensive assessment of PC, offering flexibility to users who can measure each dimension with a four-item module or choose modules based on their needs. EPAT-64 serves multiple purposes, including quality improvement and evaluation of interventions aiming to enhance PC. Its versatility empowers users in diverse healthcare settings.
- Patient-centred care
- Healthcare quality improvement
- Quality measurement
- Surveys
- Health services research
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
- Patient-centred care
- Healthcare quality improvement
- Quality measurement
- Surveys
- Health services research
WHAT IS ALREADY KNOWN ON THIS TOPIC
Patient-reported experience measures (PREMs) can be used to assess patient-centredness (PC) from patients’ perspective. The ‘Experienced Patient-Centeredness Questionnaire’ (EPAT) is the first PREM to comprehensively assess 16 dimensions of PC.
WHAT THIS STUDY ADDS
In this study, we tested the psychometric properties of all items developed for the EPAT and developed the 64-item version of the EPAT (EPAT-64), which demonstrated good psychometric properties.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The EPAT-64 can be used in research and routine care, for example, to evaluate interventions, provide feedback to healthcare professionals, support quality improvement, set benchmarks and consequently improve PC.
Background
Patient-centredness (PC) is a key element of high-quality healthcare1 2 and positively associated with several health-related outcomes.3–6
To overcome ambiguities in the definition of PC, Scholl et al developed the integrative model of PC. The model includes 16 dimensions: (1) essential characteristics of the clinician, (2) clinician–patient relationship, (3) patient as a unique person, (4) biopsychosocial perspective, (5) clinician–patient communication, (6) integration of medical and non-medical care, (7) teamwork and team building, (8) access to care, (9) coordination and continuity of care, (10) patient safety, (11) patient information, (12) patient involvement in care, (13) involvement of family and friends, (14) patient empowerment, (15) physical support and (16) emotional support. This model is based on a thorough systematic review, integrating definitions from over 400 full texts and thus providing a comprehensive model.7 Furthermore, it has been validated in several empirical studies (eg, ref 8 9). Using this model grounded our study in a broad and nuanced understanding of the various existing definitions of PC.
Effective management of PC relies on sound measurement.10 11 Patient-reported experience measures (PREMs) capture patients’ reports on whether or how often they have experienced specific processes or behaviours in their healthcare,12 and positive patient experiences are associated with several positive outcomes.13 14 While many PREMs on PC have been developed,12 15 none assessed all 16 dimensions of PC as defined by Scholl and colleagues.7 16
To bridge this gap, we developed the PREM ‘Experienced Patient-Centeredness Questionnaire’ (EPAT).17 We used a multistep process combining theory-driven approaches and the target group’s perspective to ensure high content validity of EPAT. It is a disease-generic questionnaire in German language with an outpatient and an inpatient version. A reflective measurement model was used (ie, we assumed that specific aspects of patients’ experiences are causally determined by underlying latent dimensions of PC). The preliminary EPAT version comprised 120 items for outpatients and 121 items for inpatients. Details of item development are reported elsewhere.17 However, a questionnaire of this length lacks feasibility and it was not yet psychometrically tested.
Thus, this study aimed to select items for the final 64-item version (EPAT-64) and test its psychometric properties.
Methods
Study design
Development and psychometric testing of the EPAT was part of the ASPIRED study (Assessment of Patient-Centeredness Through Patient-Reported Experience Measures), a 6-year mixed-methods study aiming to assess patients’ perspective on PC.17 18 Here, we report on the psychometric testing of EPAT with cross-sectional surveys. First, we assessed item characteristics of all items of the preliminary EPAT. Those results informed the item selection. For each of the 16 dimensions, a reduction to four items was decided, as the minimum number of items to test measurement models is 4.19 Hence, the final EPAT version has 64 items (EPAT-64). Consequently, we analysed structural validity, construct validity and reliability for the EPAT-64 items only in the same data set. As there are no guidelines for reporting psychometric evaluations for PREMs, we used the COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) guideline on reporting patient-reported outcome measures where applicable (see online supplemental appendix 1).20
Supplemental material
The EPAT questionnaire
The preliminary EPAT version and the final EPAT-64 comprise 16 domains7 8 and have an inpatient and an outpatient version each. All items are rated on a 6-point Likert scale of agreement (1=completely disagree, 2=strongly disagree, 3=somewhat disagree, 4=somewhat agree, 5=strongly agree, 6=completely agree). Every item offers the option ‘does not concern me’. For most items, a high score implies high-experienced PC, while nine items are reversed in the preliminary version and only one item in the EPAT-64. EPAT was developed and tested in German. The final EPAT-64 was translated into English using a team-based translation process consisting of translation, review, adjudication, pretesting and documentation (TRAPD).21
Data collection and participants
We included participants from June 2020 to February 2022. As planned in the protocol,18 we used consecutive sampling in 10 inpatient and 14 outpatient healthcare institutions (eg, hospitals, primary care centres) in the metropolitan area of Hamburg, Germany. Outpatients received the questionnaire before their appointment and were asked to report on their experience during the last 4 weeks within the respective outpatient clinic. Inpatients received the questionnaire on day of discharge or the day before and were instructed to reply after discharge reporting on their whole stay. As participant inclusion was slowed down by the ongoing SARS-CoV-2 pandemic, we supplemented paper-and-pencil questionnaires with an online survey using LimeSurvey (LimeSurvey, Hamburg, Germany) in December 2020. We additionally invited participants via community-based strategies (ie, social media, postings in supermarkets, self-help groups and patient organisations from all parts of Germany). They were asked to refer to their last outpatient visit or inpatient stay within the past month.
Inclusion criteria were age of 18 years or older and being currently treated for at least one of the following four medical conditions: cardiovascular diseases, cancer, musculoskeletal diseases and mental disorders. Inclusion was based on self-report. As we aimed to create a generic measure, we covered the most prevalent treatment settings and medical conditions to ensure representation of diverse healthcare experiences. We aimed to include 250 inpatients and 250 outpatients for each of the four medical conditions, resulting in a total target sample size of 2000.18 This allows for the use of maximum likelihood or weighted least squared estimators in the investigation of moderately complex models and for the detection of weak but meaningful correlations (r=0.20) with an alpha error probability of 0.05 and a power of 0.90.
In addition to EPAT, participants were asked to complete the first item of the German version of the 12-item Short Form Survey, assessing general health status.22 This item has an excellent item difficulty of 0.4 and high acceptability (less than 4% missing values). We assessed satisfaction with care with the German version of the 8-item Client Satisfaction Questionnaire,23 a unidimensional questionnaire with high internal consistency (Cronbach’s alpha=0.9). We also administered the European Health Literacy Questionnaire (HLS-EU-Q16),24 a short version of the HLS-EU-Q47 with low ceiling effects, strong correlation with the long version (r=0.82) and high internal consistency (Cronbach’s alpha=0.90).25 The HLS-EU-Q16 can assess overall health literacy with a sum score of all items. Furthermore, we asked about demographic factors (eg, gender and age) and health status (eg, years since receiving the corresponding diagnosis). Finally, first in the online version and from June 2021 in the paper questionnaires, we included two control items (‘To show that you are reading attentively, please check ‘somewhat disagree’/‘somewhat agree’.’). Data collection was anonymous. All participants gave written consent. Participants sent back paper questionnaires free of charge and had the possibility to receive an incentive of €10 on completion of the questionnaire.
Data analyses
First, we analysed all items of the preliminary EPAT version. We then selected items for EPAT-64 in a team discussion with five experts in the field of PC (IS, LK, MH, JMZ, EC). Three of them (IS, JMZ, EC) developed EPAT and were familiar with the qualitative data used for development.17 For each dimension, we chose four items resulting in 64 items covering 16 dimensions of PC. We considered the following psychometric properties during item selection: (1) percentage of missing values as proxy for acceptance,26 (2) percentage of patients replying ‘does not concern me’ as proxy for relevance, (3) item difficulty calculated by standardising the mean to range from 0 to 1 (recommended range 0.2–0.8),27 (4) within-dimension interitem correlations (recommended range >0.3),28 (5) within-dimension corrected item–total correlation (recommended range >0.3)28 and (6) content validity (eg, how important an item is to the definition of the given dimension).
For EPAT-64, we examined structural validity. We had a clear hypothesis concerning the number and content of the dimensions to be measured, based on substantial previous work.7–9 17 Therefore, we used confirmatory factor analysis (CFA)29–31 using a robust version of maximum likelihood estimator to assess five different models: (1) unidimensional model, (2) correlated first-order dimension model, (3) hierarchical model, (4) bifactor model with uncorrelated dimensions and (5) bifactor model with correlated dimensions. See online supplemental appendix 2 for detailed descriptions and depictions of all models. For examination of model fit and selection, we followed recommendations by using a combination of model fit indices32: root mean square error of approximation, standardised root mean square residual, Tucker-Lewis Index and Comparative Fit Index. For comparison of models, we used Akaike information criterion and Bayesian information criterion. Further, we report the χ2 test statistic of model fit.33 For the selected model, we report standardised factor loadings and average variance explained (AVE). We planned to test this model in the four medical conditions to examine measurement invariance between groups.28 Further, we examined reliability by McDonald’s Omega hierarchical which allows to estimate reliability in models with a bifactor structure.34
Supplemental material
Construct validity was assessed by correlating the model-based factor scores of each dimension with two other measures as described in the study protocol.18 Discriminant validity was tested by examining correlations with general health status.22 We hypothesised that the magnitude of these correlations should be below 0.3.35 For convergent validity, we used correlations with satisfaction with care,23 expecting correlations to exceed 0.5.35
We used SPSS V.27.0 (IBM) to enter paper-and-pencil questionnaire data, to check their validity by a second person, to import data from LimeSurvey and to calculate sample characteristics. All other data cleaning and analyses were done with R V.4.3.2 (R Core Team, Vienna, Austria), specifically we used the lavaan package36 for CFA, BifactorIndicesCalculator37 for Omega hierarchical and psych package38 for all other psychometric analyses. We excluded participants if they responded to less than 70% of EPAT items or if they failed both control items. Where applicable, we used full information maximum likelihood to deal with missing values,39 treating ‘does not concern me’ replies as missing values. We report results for all medical conditions combined, but analysed data separately for inpatient and outpatient settings.
Results
Sample
For the paper-and-pencil version, 4788 outpatients were invited to participate, of whom 2357 (49.2%) gave consent and 905 (18.9%) returned the questionnaire. Further, we invited 3046 inpatients, of whom 1799 (59.1%) gave consent and 704 (23.1%) returned the questionnaire. For the online questionnaire, 1042 patients consented and 458 patients replied to at least 70% of the items (44.0%). We have no information on patients who denied consent or dropped out to perform dropout analyses. In total, questionnaires of 1092 outpatients and 931 inpatients were included in our analyses (see the flow chart in figure 1).
Participant flow chart. EPAT, Experienced Patient-Centeredness Questionnaire.
Of the included outpatients, 59.1% were female, 38.5% were male and 0.5% identified as diverse. Mean age was 52.8 years (SD=17.4, range 18–95). Mean time since initial diagnosis of the respective disease was 11.4 years (SD=11.0). Of the inpatients, 41.3% were female, 55.5% were male and 0.6% identified as diverse. Mean age was 55.5 years (SD=17.4, range 18–92). Mean time since initial diagnosis was 8.5 years (SD=10.0). See table 1 for further details on demographic and health-related variables. For subgroup sample characteristics regarding medical conditions refer to online supplemental appendix 3. For a comparison of paper-and-pencil and online questionnaires please see online supplemental appendix 4.
Supplemental material
Supplemental material
Sample characteristics
Item characteristics
Results on all items of the preliminary EPAT, which were the basis for item selection of EPAT-64, are reported in online supplemental appendices 5 and 6 for inpatients and outpatients, respectively.
Supplemental material
Supplemental material
For outpatients, all 64 selected items showed high acceptability (highest percentage of missing responses was 2.7%). Regarding relevance, 47 items had less than 30% of participants answering ‘does not concern me’ and seven items had over 50%. While 46 items showed excellent item difficulty (between 0.2 and 0.8), 18 items showed ceiling effects (item difficulty >0.8). All but one item had an item–total correlation greater than 0.3.
For inpatients, again, all 64 items had high acceptability (highest percentage 2.7%). We included 55 items with a ‘does not concern me’ rate below 30% and two items above 50%. Item difficulty was excellent for 42 items, 22 items showed ceiling effects. All items had item–total correlation greater than 0.3.
Item characteristics of all items included in the EPAT-64 are shown in table 2. For item wording, please refer to the full questionnaires at www.uke.de/epat.
Item characteristics per medical condition are shown in online supplemental appendix 7 (outpatients) and online supplemental appendix 8 (inpatients).
Supplemental material
Supplemental material
Item characteristics of the EPAT-64
Structural validity
Model fit indices for the five tested models are shown in table 3.
Model fit indices for confirmatory factor analysis on EPAT-64
The best fit was found for the bifactor model with correlated dimensions. Here, all items loaded on their respective specific dimension and on a general dimension. Specific dimensions were intercorrelated, but their correlation with the general factor was restricted to be zero. Standardised factor loadings and AVE of this model are reported in table 4. Unstandardised factor loadings and error variances are shown in online supplemental appendix 2. Correlations between specific dimensions showed a wide variability ranging from −0.140 to 0.824 (see online supplemental appendix 9).
Supplemental material
Standardised factor loadings, average variance explained and reliability coefficients of the correlated bifactor model
In the inpatient and outpatient samples, standardised loadings for the general factor were highest for the ‘Essential characteristics of the clinicians’ dimension and lowest for ‘Integration of medical and non-medical care’. On their specific factors, these two dimensions had the lowest and highest standardised loadings, respectively. The general factor AVE was highest for ‘Clinician-patient relationship’ for outpatients (41.0%) and ‘Coordination and continuity of care’ for inpatients (36.8%). Specific factors showed the highest AVE for ‘Essential characteristics of the clinicians’ in outpatients (87.9%) and for ‘Patient safety’ in inpatients (97.2%).
We further explored the general factor in the bifactor model by calculating correlations with other measures. We found a strong correlation between the model-based scores of the general factor and satisfaction with care (outpatients: 0.73, inpatients: 0.83). Correlations of the general factor with health literacy, general health status, comorbidity, age and time since initial diagnosis were all below 0.3.
We were not able to examine measurement invariance across medical conditions. The combination of small sample sizes within condition subgroups, high rates of missing values due to ‘does not concern me’ and the complexity of the tested model led to estimation problems (eg, non-convergent models, not positive definite variance-covariance matrix of estimated parameters).
Reliability
McDonald’s Omega hierarchical was high for the general factor (outpatients: 0.868, inpatients: 0.908) and varied widely for the specific dimensions (outpatients: 0.214–0.872, inpatients: 0.128–0.847). To support interpretability of sum scores for specific dimensions, we examined their correlations with the model-based factor scores (online supplemental appendix 10). Results suggest that these sum scores measure a mixture of the general factor and the specific dimension.
Supplemental material
Construct validity
Using model-based factor scores, discriminant validity was confirmed for all specific dimensions and the general factor (correlations with general health status <0.3). Convergent validity was confirmed for the general factor, which was strongly correlated with satisfaction with care, while the specific dimensions showed no strong correlations. Correlations for all dimensions and the general factor are shown in online supplemental appendix 11.
Supplemental material
Discussion
Summary of the findings
EPAT-64 is a PREM to assess 16 dimensions of PC from patients’ perspective. All items demonstrated high acceptability. Most items had low rates of patients answering ‘does not concern me’. Regarding item difficulty, most items were within the recommended range, but about 30% of the items showed ceiling effects. Item–total correlations were good for all but one item. CFA showed the best fit for a bifactor model. Discriminant validity was confirmed for all dimensions and the general factor. Convergent validity was confirmed for the general factor only. McDonald’s Omega hierarchical suggests good reliability of the general factor, while there was wide variability for the specific dimensions.
Strengths and limitations
EPAT-64 is the first PREM comprehensively assessing PC.15 16 One of its strengths is the thorough development and item selection process, which prioritised content validity.7 8 17 Some items of EPAT-64 showed suboptimal item characteristics, in particular ceiling effects, which may limit the ability to discriminate between high and low PC. These items were included nevertheless as we selected items based on several quantitative criteria and content. Further, we decided to choose four items per dimension, as we aimed for parsimony while requiring at least four items for statistical model identifiability.19 For dimensions that were covered by only four items in the preliminary EPAT,17 we had no choice between items. Deleting dimensions was not an option, as all dimensions are relevant to patients.8 We argue that selecting items solely based on quantitative data may lead to more reliable questionnaires, but risks compromising content validity.19 In particular in outpatients, some items had high rates of ‘does not concern me’ answers, especially those beginning with ‘if’ or ‘when’ conditions. For instance, 54.7% of the outpatient sample replied this way to the item ‘When I had pain, I was helped quickly’. However, this item is essential for the definition of the ‘Physical support’ dimension8 and highly relevant for everyone with pain. Overall, some items being less relevant to certain patient groups influences EPAT-64’s generic usability.
CFA showed best fit for a bifactor model with correlated dimensions. This model postulates that all items load on a general factor and on their specific dimension, which is independent of the general factor. How well items’ variance was explained by their specific and general factors and specific factors’ reliability varied. Thus, not all items are well represented by the bifactor model. We decided against further modifying this model, as factor structure decisions should combine data and theory-driven approaches to avoid overfitting and ensure meaningful interpretation.7–9 17 Overall, the bifactor model showed good results for global model fit indices, reflecting our underlying model of PC, but the local fit for specific items needs further investigation. An advantage of the bifactor model is that it enables users to partition scores into an estimation of the general factor and of the specific dimension.40 The general factor likely captures more subjective aspects of PC, as it strongly correlates with satisfaction with care (see convergent validity). The specific dimensions of PC seem to explain variance not captured by the general factor, confirming the underlying conceptual model. We assume that these dimensions capture more objective aspects of PC (ie, whether something happened rather than how well it went). No strong correlations between their factor scores and satisfaction also indicate this (see convergent validity). This differentiation is important in light of definitions of PREMs12 which stress patient experience and satisfaction being two distinct constructs.41 Satisfaction is subjective and influenced by individual expectations, patient experience is more objective and focuses on whether certain processes occurred.12 While EPAT-64 was designed as a PREM, the general factor reflects satisfaction. Since experiences and satisfaction are ‘complexly related’41 and all self-reports introduce some subjectivity,42 PREMs likely correlate with satisfaction. The EPAT-64 bifactor model may allow statistical differentiation with the general factor likely relating to satisfaction and the specific dimensions likely capturing experiences. Thus, we recommend users to apply the whole EPAT-64. This way, they can model the bifactor model in their data and calculate both general and dimension scores. Yet, in routine care, there might be sparse resources for administering a 64-item questionnaire or insufficient sample size for estimating bifactor models. Hence, choosing the most relevant dimensions for a context and calculating sum scores per dimension might be more feasible. This flexibility allows adapting the questionnaire to specific questions, areas of application and resources. Missing values should best be dealt with using multiple imputation or full information maximum likelihood.43 In routine care, when sample size and resources are limited, one missing value per person and dimension can be replaced by the average of the other three items of that person.43 For persons with more missings, no sum score for the respective dimension should be calculated. When interpreting those sum scores, one should consider that this score combines subjective and objective PC assessments. Further, the specific dimensions are differently reliable and correlations between them make interpretation of sum scores more challenging. Online supplemental appendix 10 helps guide sum score interpretation, but more research is needed.
As a limitation, we were not able to examine measurement invariance across different medical conditions. To make transferability across medical conditions likely, we developed items generically and tested them in four medical conditions that affect a large proportion of the population. Yet, due to the complex measurement model and the large number of ‘does not concern me’ replies in some items, we were not able to confirm a common factor structure across conditions. Further, the high dropout rate after consent is a limitation. As self-selection bias is possible, acceptance and feasibility of EPAT-64 data collection need further testing. The ongoing SARS-CoV-2 pandemic might have influenced our results, in particular, dimensions like ‘Involvement of family and friends’ or ‘Access to care’. Furthermore, pandemic access restrictions led us to adding an online survey. While this broadened the sample and might have improved generalisability, there might be differences between online and paper-and-pencil subgroups. Paper-and-pencil participants were treated in or near Hamburg and received the questionnaire on the day of their treatment; online participants came from across Germany and might have responded much later. Further investigation of influences of data collection modality on EPAT-64 is needed. Notably, since we aimed to design a PREM, we defined PC based on interactions between patients and healthcare professionals. Other models view PC more broadly and include aspects that occur ‘further away’ from the care system (eg, physical activities and self-management practices).
Implications
There is a wide range of possible applications of EPAT-64. Measuring PC is essential for research on the current state of PC and on fostering PC. Internationally, PREMs are used to provide feedback to healthcare providers, inform quality improvement, public reporting, benchmarking and value-based purchasing.44–46 To improve patient experiences, data from PREMs need to be used in a coordinated approach.47 Emerging evidence suggests that being informed about patient experiences is associated with improved communication,48 and meetings to discuss PREM results are associated with improved patient experience.45
Psychometric testing in independent samples is indispensable to ensure that the psychometric properties found here hold in other populations and contexts. Particular attention should be paid to the replication of the bifactor structure with correlated dimensions and measurement invariance. In addition to validity and reliability, aspects of fairness and responsiveness should be investigated.49 Regarding testing in other patient groups, the EPAT-64 is currently adapted to assess PC in the context of healthcare and support services for women with unintended pregnancy.50 Further adaptations to other healthcare contexts, such as rehabilitation facilities or non-academic hospitals, could be considered.
Internationally, no comparable PREM that comprehensively assesses all dimensions of PC exists.15 Therefore, cross-cultural adaptations and translations of the EPAT-64 could provide an opportunity to promote PC around the world. In addition, translations would help reach people in Germany not fluent in German. An English EPAT-64 translation that used the TRAPD process21 can be downloaded from www.uke.de/epat-en. However, further tests on comprehensibility and psychometric properties are required before it can be used.
Conclusion
The EPAT-64, a novel questionnaire assessing 16 dimensions of PC from the patient’s perspective, emerges from a meticulous development process that blends theory-driven with data-driven decisions. Psychometric testing within a substantial patient cohort underscores its robust psychometric properties. Nevertheless, further research is imperative to refine the interpretation of scores. The adaptability of this tool to specific purposes positions it as a valuable instrument for both assessment and cultivation of PC.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by the Ethics Committee of the Medical Association Hamburg (study ID: PV5724). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We thank our student assistant Sibel Gürü and our research interns Lucas Bolz, Antonia Buchal, Sophia Ehrentraut, Carolin Guddat, Zoe Henning, Constanze Mahn, Yannick Mosemann, Philip Niemann, Svea-Marie Nohrden, Paul Packheiser, Ronja Platzeck, Mareike Remmers, Ramona Rusch, Laura Stefan, Melvin Schneider and Lisa Strelow for carrying out data collection. Further, we thank all our collaboration partners who gave us access to their patients for the data collection: Klinik für Psychosomatische Medizin und Psychotherapie (UKE) III, Medizinische Klinik und Poliklinik (UKE), Strahlentherapie und Radioonkologie (UKE), Martini-Klinik, Evangelisches Krankenhaus Alsterdorf, Verhaltenstherapie Falkenried, Bundeswehrkrankenhaus Hamburg, Asklepios Westklinikum Hamburg, Psychiatrische Ambulanz Dr Guth and Fachzentrum Altona für Psychiatrie und Psychotherapie.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
LK and IS are joint senior authors.
Contributors IS was the responsible principal investigator of the study and is the guarantor. IS, LK and MH were involved in planning and preparation of the study. SZ, AS, HF and EC recruited the participants and collected the data. EC analysed the data with supervision from LK. EC, SZ, JMZ, IS, MH, PH and LK interpreted the results and discussed the item selection. EC wrote the first draft of the manuscript. CB, JG, CG, CM, VM, KS-B and AS actively supported the data collection. All authors critically revised the manuscript for important intellectual content. All authors gave final approval of the version to be published and agreed to be accountable for the work.
Funding This study was funded by the German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung – BMBF) with the grant number 01GY1614. The funder had no role in decision to publish or preparation of the manuscript.
Competing interests PH is a board member of the International Shared Decision Making Society, a charitable scientific society. IS received honoraria for presentations and speeches on patient-centred care from the following commercial entities: onkowissen.de GmbH and ClinSol GmbH & Co KG.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.