Development and empirical testing of a questionnaire-based algorithm to evaluate physical and psychosocial health status in performing artists / Entwicklung und empirische Testung eines fragebogenbasierten Algorithmus zur Erhebung des körperlichen und psychosozialen Gesundheitszustands darstellender Künstlerinnen und Künstler

The aims of the present study were (1) to compose an algorithm including standardized validated patient-related outcome measures (PROMs) for the evaluation of the individual physical and psychosocial health status of performing artists with musculoskeletal health complaints, taking into account its complexity and multidimensionality, and (2) to empirically test the application of the proposed algorithm in a broad sample of performing artists.

Questions on the individual characteristics of the patient like sociodemographic information as well as performing-art-specific information including the professional level were compiled in a self-administered General Questionnaire. The relevant aspects were selected on the basis of Dommerholt (2009) and were supplemented by aspects described in established performing-artists-specific literature (Ackermann & Adams, 2004; Sataloff et al., 2010; Chan & Ackermann, 2014; Ackermann et al., 2011).

Additionally, PROMs relevant for the performing-art-specific history taking were selected based on the literature search and following criteria, which were defined a priori by the authors. The included PROMs should

be performing-artist-specific,

Regarding the non-performing-art-specific questionnaires covering the psychosocial dimension and addressing the different body regions, it was decided to include well-established PROMs commonly used in the physiotherapeutic context, which cover the aspects mentioned above.

Regarding the performing-artist-specific PROMs, the database PubMed was searched for full texts in English or German using the terms

(instrumentalist* OR musician*) or (sing* OR vocal* OR voice) or (danc*)

AND (“patient related outcome measure” OR self-report OR questionnaire OR index OR inventory OR survey), respectively, in title/abstract.

Results were screened for relevant articles matching the above-mentioned criteria. Whenever an appropriate PROM was found, a transcultural adapted German version was searched. All selected PROMs are shown in table 1 (results section).

After the selection of the questionnaires, the order of the questionnaires was determined and the algorithm was defined. The fundamental idea of the algorithm was that the patient should be guided individually through a series of PROMs in order to address the individual health problem. Key questions were used at various points in this algorithm in order to decide which questionnaire or item will be displayed next. These optional questionnaires should address in depth either specifically one of the performing arts or musculoskeletal problems in a specific body region. Therefore, conditions were defined under which the participant was forwarded to specific questionnaires or items while other questionnaires or items were skipped.

The items of all questionnaires and the corresponding conditions were stored in German in the online survey tool LimeSurvey. The structure of the algorithm with the various redirects as well as spelling and grammar were pre-tested with 17 non-performing-artists. Corrections were incorporated and modifications were refined before the empirical test phase within the target group of performing artists.

The survey was applied to instrumentalists, singers, dancers, and actors. It is not only for professionally trained artists that performing can have a major impact on their quality of life, but also for amateurs and non-professionals. Conversely, regardless of the level of artistic training, discomfort, and complaints can lead to significant impairment in the practice of the performing arts and, depending on the importance of the arts to one's life, can also affect quality of life. Therefore, performing artists of all levels of professionalism/artistic training and of all genres were addressed. The survey was conducted within the region of Osnabrueck-Emsland. The aim was to get the largest possible sample so that each PROM would be completed at least once. Therefore, a snowball effect was intended and the artists were invited to forward the access to the survey to other artists as well. Internet searches were conducted for single artists of all 4 groups of performing artists as well as bands, music and dance sports clubs, choirs, theatres, etc. In a first wave, 953 performing artists were informed via email or mail about the background and aims of the project and invited to participate in the survey. In a second wave of recruitment, a reminder was sent and additional 72 instrumentalists', singers', dancers' and actors' associations, universities and other regional multipliers were informed and invited to participate or forward the invitation to their members. A link or a QR code was included in the e-mails/mails forwarding directly to the online survey.

Performing artists were surveyed from july 2019 to february 2020 in an exploratory observational design with one measurement point. The algorithm guided the study participants individually through the composition of PROMs depending on their answers to specific key questions (see Tab. 1). After stopping the online survey, the procedure of data preparation, scoring and interpretation was tested (see “statistical considerations”). One of the participants was selected by the authors, who had completed all PROMs in his/her algorithm-driven path through the survey without dropping out. His/her individual path through the survey was illustrated. The individual scores of his/her combination of PROMs were compiled in diagrams for both the individual physical and psychosocial health dimension.

For step 2 a data and a syntax file were extracted from the online tool LimeSurvey and imported to SPSS (IBM SPSS Statistics, Version 26) in order to test further analysis. For the description of the study population descriptive statistics such as mean with standard deviation, median with an interquartile range, and a range with a minimum and maximum were calculated for sociodemographic and performing-art-specific parameters extracted from the General Questionnaire. The PROMs were scored including preparation like recoding, reversed scaling, etc. as indicated by the respective authors (Tab. 1). In order to improve comparability, all PROM scales were normalized to a uniform 0–100 scale with lower scores indicating better outcome/less impairment. Therefore, the scales of the DASS, the K-MPAI-26, the Pain Intensity and Interference Scale of the MPIIQM-G, the NDI and the SVHI were normalized to a 0–100 scale and the scaling of the SF-12 was reversed from “higher scores indicate better outcome” to “lower scores indicate better outcome”. Cut-off values and grading were used for the interpretation of the PROM scores whenever available (in the presented example: SF-12, NDI and SVHI) (Tab. 1).

The present study was conducted in accordance to the ethical guidelines of the Osnabrueck University of Applied Sciences referring to the Guidelines for Safeguarding Good Research Practice by the DFG (2019) and the World Medical Association Declaration of Helsinki (WMA, 2013). Participants were informed about the background and aims of the study as well as contact persons and responsibilities during recruitment and again at the beginning of the survey. The first item of the online survey referred to the protection of the data to be collected and the possibility of its withdrawal at any time. If the participant gave informed consent, he or she was directed to further items. If he or she disagreed with the guidelines, the survey was stopped at that point and no data were stored. Before the analysis of the data of those who gave informed consent, personal data like email or names were removed from the data set.

An overview and descriptions of the selected questionnaires as well as the algorithm is shown in Table 1 in the supplementary material. The table includes the description of the selected PROMs clustered by health dimension and specification; the outcomes; further literature on development and psychometric properties of the original version; the availability and psychometric properties of the German versio; information on scoring and interpretation of scores; and indication of whether the questionnaire is mandatory or optional including the condition to be forwarded to the respective questionnaire (ADL = activities of daily living), and the adjustments made for the needs of the present study).

At the very beginning of the survey the participant was informed about the project, ethical considerations, data protection, and the possibility of withdrawal at any time. The online survey started with the General Questionnaire only if the participant gave informed consent.

The General Questionnaire was self-administered by the authors of the present study in order to assess sociodemographic as well as performing-art-specific characteristics which are relevant for further interview-based history taking. Additionally, items on general information in the MPIIQM-G, SVHI, VHI, and the MHQ were removed from these questionnaires and summarized in the General Questionnaire.

The SF-12, DASS-21, and K-MPAI-26 covered the psychosocial health dimension, all other questionnaires covered the physical health dimension. The K-MPAI, all three versions of the MPIIQM-G, the SVHI, and the VHI were selected as performing-art-specific PROMs. The MPIIQM was originally developed especially for orchestra musicians (Berque et al., 2014). However, in the present study, it was used for all types of instrumentalists and—in the reworded versions—also for singers/actors and dancers in order to compare the musculoskeletal complaints between groups. The SVHI and VHI were selected as voice-specific PROMs. In the present study, the VHI was used to evaluate voice-specific complaints in actors, even though it was originally developed for every type of voice patients, not especially for actors (Jacobson et al., 1997). The CF-PDI, NDI, QuickDASH, MHQ, ODI, HOOS-PS, KOOS-PS and the FAOS were selected as the body-region-specific PROMs.

A psychometrically tested German version was available for the DASS (Nilges & Essau, 2015), the NDI (Cramer et al., 2014), the QuickDASH (Institute for Work&Health, 2006a+b), the ODI (Mannion et al., 2006a+b), the HOOS-PS and the KOOS-PS (Roos, no year), and the FAOS (van Bergen et al., 2014). The General Questionnaire was not validated, because it was not used to gain information on a certain construct. Kenny et al. (2004) tested the original English 26-item-version of the K-MPAI for internal reliability and validity. A German version of the K-MPAI-26 was psychometrically tested for validity and reliability by Heimann and Kummermehr (2020) but has not yet been published in a peer-reviewed journal. The MPIIQM-G was validated in the German language by Möller et al. (2018). In the present study, it was slightly reworded use regarding singers/actors and dancers in order to assess the musculoskeletal problems in a performing-arts-specific way. However, the reworded versions were not tested psychometrically before use. The English version of the CF-PDI had already been validated at the time of questionnaire selection (La Touche et al., 2014). The German cross-cultural adapted translation (von Piekartz et al., 2021) was performed before questionnaire selection in the present study and was used with permission, even though the psychometric evaluation had not been completed. In the standardized translated and validated German consensus version of the MHQ by Knobloch et al. (2011c) no distinction between right and left hand was made in the subdimension of “pain”. Therefore, in the present study, the answers given in this subdimension were indicated to be currently affected the most in order to be able to apply the MHQ scoring algorithm provided by the licensors. The evaluation of the secondary disease was self-administered and not psychometrically tested. The questions were based on the standardized German Pain-Questionnaire (Deutscher Schmerzfragebogen) (DGSS, 2018).

The performing artist was guided individually through the survey and was directed to PROMs, which address his or her complaints more in-depth. Irrelevant questions for the respective participant were skipped based on specific key questions. Depending on the indication of the main performing art practiced (information provided in the General Questionnaire) the participant was directed to one of the versions of the MPIIQM-G and to one of the voice-specific questionnaires if the main performing art was singing or acting, whereas these PROMs were skipped for instrumentalists and dancers. Furthermore, musculoskeletal problems were surveyed in the MPIIQM-G, but they were addressed in depth by one of the body-region-specific PROMs. The MPIIQM-G is constructed in such a way that the participant is asked about the prevalence of musculoskeletal problems within 4 different time slots (ever, during the last 12 months, during the last month/4 weeks, currently/in the past 7 days). If the participant indicates having musculoskeletal pain or problems within the last month and/or currently, he or she is directed to a body chart, in which all complaints as well as the one area of worst pain should be named (Möller et al. 2018). In the present survey, depending on the location of this main pain prioritized by the participant (MPIIQM-G, item no. 13b), the survey algorithm led to one of the PROMs, that specifically addressed complaints in that respective body region. For example, an artist, who indicated his/her main pain in the neck, was directed to the Neck Disability Index (NDI), the one, who reported his/her main pain in the lumbar spine, was forwarded to the Oswestry Disability Index (ODI) (Fig. 2).

In step 1, PROMs in German covering both the physical and psychosocial health dimensions were selected based on a literature search in order to determine the self-perceived health status of performing artists as comprehensively as possible and to provide a basis for further interview-based history taking. The focus was on musculoskeletal complaints and the associated individual risk factors that may contribute to their development or maintenance. The psychosocial health dimension was measured by the SF-12, the DASS, and the K-MPAI-26. The physical health dimension focusing on voice-specific complaints was measured by the SVHI or the VHI. The physical health dimension focusing on musculoskeletal complaints was measured with the MPIIQM-G and the 8 body-region-specific questionnaires CF-PDI, NDI, QuickDASH, MHQ, ODI, HOOS-PS, KOOS-PS, and FAOS. With the help of a newly composed algorithm, the individual performing artist is offered only those questionnaires that are relevant to him/her, skipping all other questions.

With the selected PROMs, the algorithm covers all criteria that were determined a priori by the authors (cf. result section, step 1). Providing the survey in online format allowed automatic skipping of those questionnaires that are not relevant for the individual participant and therefore allowed a specifically tailored questionnaire-based history taking. The pre-test supported the technical function of the algorithm before it was applied to the target group.

Regarding psychometric properties, it should be taken into account that the validation of the German versions of the MPIIQM-G, which has been reformulated by the authors for the groups of singers, dancers and actors, respectively, has not yet been carried out. Therefore, the psychometric quality of the algorithm cannot yet be clearly determined.

In step 2, the algorithm was applied to performing artists in order to empirically test it within the target group and within subjects with “real” performing-arts-associated complaints. Each PROM was completed by at least one subject (cf. Fig. 1&2) as intended. For one individual subject (P0014) who was exemplarily selected by the authors, the data from all PROMs that resulted from his individual path through the survey, were analyzed (scores were calculated and graphically presented). This exemplifies what information on both the physical and psychosocial health dimension of a performing artist the algorithm could provide in order to support the physiotherapist in further patient management. The evaluation and graphical representation of the results have an interesting added value for further history taking in the personal interview. At this point in time, the creation of these diagrams and interpretation of scores is rather time consuming and should be automated for the implementation in the common physiotherapy practice.

The respective participant was asked to fill various PROMs (ranging from 6 to 8) in order to cover both the physical and psychosocial dimension of health at a high level of detail. In order to reduce client's burden, it is recommended to use validated short versions of the PROMs.

In the present study, a high level of detail in evaluating both the physical and psychosocial health status of performing artists was intended. This was achieved by addressing the different possible localizations of the individual major problem as well as artist-specific complaints and both physical and psychosocial individual risk factors through the different PROMs. Additionally, the sample covered a wide range of performing artists from different artistic disciplines and level of experience and reported on “real” performing-arts-associated complaints as intended. The online format enabled the participant to complete the survey independently of the restricted time with a therapist. The physical and psychosocial health status of an individual performing artist presented in its complexity and multidimensionality represents a therapeutically important basis for comprehensive patient management. As soon as there are relevant PROMs with a higher psychometric quality, these should be integrated in the algorithm. That might improve the psychometric quality of the proposed algorithm to provide valid and reliable information on the individual's physical and psychosocial health status and improve clinical benefits. However, detailed questioning with the proposed algorithm as well as graphical representation of the results might provide important information as a basis for further physiotherapeutic patient management.

For an application in physiotherapeutic practice, the use of short versions is recommended in order to save time when filling the algorithm. An automation of provision and interpretation of the information resulting from the survey would be a helpful and timesaving step in the implementation to the physiotherapy practice. In future research, performing artists should be asked to provide feedback on the usability of the algorithm.

In addition, physiotherapists would need instruction on scoring and interpretation of PROM scores in order to reliably use the algorithm and PROMs. The clinical benefits for individually tailored patient management should be evaluated based on the feedback of physiotherapists using the algorithm for their therapeutic work.

For the clinical use of the algorithm, the evaluation of changes of complaints over time would be of special interest for both the therapist and the patient. Therefore, the algorithm, or at least selected PROMs that have revealed problems in one of the areas, have to be used repeatedly. In that case, additional values like the minimal detectable change (MDC) or minimal clinical important change (MCIC) are important for clinical interpretation. The benefits of the proposed algorithm for better artist-specific outcomes over time should be evaluated, as well.