The study will take place at the University Medical Center Groningen (UMCG), which has already hosted the Meaningful Music in Healthcare (MiMiC) live intervention [25]. The intervention study follows an on-off nonrandomized controlled trial (NCT) design [26]. The NCT is necessary for this particular intervention in practice because the live music intervention cannot easily be randomized to skip patients in the same ward without them hearing the music or seeing the musicians as they play for other patients and nurses.

In this NCT design, treatments and data collection will be conducted in 15 blocks over the course of 4 years. According to the 29-day treatment block, the three 5-day treatments (live music intervention, recorded music active control intervention, and standard care control) will be offered with a 7-day washout between treatments. The washout will minimize the number of participants whose hospital stay overlaps with another treatment. Based on the previous research (UMCG research register 201600541), we expect approximately 8 full data sets per 5-day treatment block.

Note that while the intended outcome of the music intervention is therapeutic (eg, experiencing reduced pain), the music intervention is conducted by conservatory-trained music practitioners but not by music therapists. Music therapists are licensed therapists who have a degree in music therapy including clinical training. The music practitioners in this study are primarily professional musicians who have taken 1 music-conservatory course and several internships to train them for clinical settings but they are not therapists.

Design procedures are chronologically displayed in Figure 1. Preoperative questionnaires will be collected to assess cognitive function (6-item cognitive assessment [6CIT]) [27,28] and musical background (excerpt from the Dutch Musical Background Questionnaire [DMBQ]) [29,30]. Postoperative data collection procedures per day of treatment include pain visual analog scale (VAS) ratings, heart rate (HR), and blood pressure measuring 4 times a day (before, directly after, 3-hour, and 6-hour postintervention). EEG electrodes will be applied once in the morning and 4 EEG recordings will be collected with VAS ratings, HR, and blood pressure (before, during, 3-hour, and 6-hour postintervention). A blood sample will be collected once daily and integrated with standard care as far as possible (eg, if blood is drawn as part of standard care, the sample will be collected from the same needle). A schedule of assessments may be found in Table 1. Daily dosage of pain medication as well as any incidence of delirium (delirium observation screening scale [DOSS]) [31] or POCD will be collected.

The study will include patients who have undergone elective surgery in the UMCG and are hospitalized for at least 1 day following surgery. All patients must be 18 years or older and able to make their own decisions. Patients must further be able to hear music with normal or corrected normal hearing (ie, a hearing aid).

In order to be eligible to participate in this study, a subject must meet all of the following criteria: (1) 18 years or older, (2) postsurgery hospitalization (all surgery types), (3) elective surgery, (4) able to hear music, (5) able to give informed consent, and (6) able to communicate.

Patients will be informed as soon as the date of hospital admission is set in the weeks that the study is planned. The patient will receive an information letter and informed consent from the provider or it will be sent to them by email. A member of the research team will contact the patients after a minimum of 24 hours for participation and answer possible questions. Patients will have up to the time of their surgery to decide, allowing for time to fill out the preoperative questionnaires. Consent can be withdrawn at any time at which point all data collection and treatment will stop.

This protocol has been approved by the Central Medical Ethics Committee of the UMCG (METc2021.359) and is registered with the Netherlands central commission of human research (CCMO-register number NL76900.042.21). The informed consent letter is written in plain, nonacademic Dutch with no medical jargon. Only pseudoanonymized (deidentified) data will be published. Patients participating in the study will receive a small token gift (such as a hand lotion or lip balm).

This study will take the form of an on-off NCT [26]. An on-off NCT takes into account real-world setting limitations of certain intervention types and allocates participant inclusion by alternation. The main reason for this is that MiMiC is a well-researched intervention, whose practical implementation is not suited for randomly selected participants to not participate. On the one hand, musicians will visit the nurses’ station and play music in various rooms around the ward; thus, it is not possible to prevent participants who may be allocated non-MiMiC treatment from hearing the live music. On the other hand, a central tenet of the MiMiC intervention is inclusion and participation, thus “rejecting” patients who would want to participate in music they hear on the ward, on the basis of trial group allocation, goes against the spirit of the intervention.

In contrast, the NCT relies on the randomness of who is scheduled for surgery during a given intervention week but does not exclude any patient from the activity being conducted (listening to live or recorded music). Therefore, the treatment schedule (live music intervention, recorded music active control, and standard care control) will be predetermined months in advance, and the patients included will be offered whichever treatment is occurring during their hospital stay. The patients will, however, be asked for their consent before they are informed which treatment will be offered during their surgery recovery. This will attempt to prevent patients who are initially biased for inclusion based on their affinity to music or not.

After patients are recruited to the study, they will be allotted to the treatment occurring during the scheduled postoperative hospital stay. Noninvasive procedures include preoperative cognitive testing (6CIT), postoperative VAS pain ratings, HR, blood pressure, respiration rate, and EEG recording. Invasive procedures include obtaining blood samples, after which the serum will be analyzed in a laboratory for neuroinflammatory markers IL6, NGAL, CRP, and serum TPSO. An example schedule of procedures can be found in Table 1.

The main end point, VAS pain ratings, requires patients to subjectively rate their pain on a plastic VAS where they move a tick mark to the desired location on a 0-10 scale [1].

The secondary end points, HR and blood pressure, will be collected from standard equipment on the ward at the time of obtaining VAS pain ratings.

A short, 6-item questionnaire [27,28] will be added to the standard preoperative intake in order to assess preoperative cognitive status, as well as a short questionnaire about the musical background (excerpt from the DMBQ) [29,30]. Both the cognitive status and the musical background will help to interpret the results. The preoperative cognitive status will serve as a baseline for postoperative cognitive status as well as alpha power in EEG recordings. The musical background will be assessed as a possible explanation for the success of the live or recorded music interventions. Other complications (classified according to Clavien-Dindo [33], nursing-sensitive outcomes such as dehydration and falling [34]) will be explored for changes in the music groups compared to care-as-usual to derive hypotheses for future studies.

The primary outcomes of this study are VAS subject pain ratings. The F test is described immediately below. Statistics will be conducted with SPSS (IBM Corp) or equivalent software [36].

Pain ratings will be collected 4 times a day with a VAS. VAS mean and SD will be reported, and box plots, histograms as well as P values for Kolmogorov-Smirnov tests will be evaluated for normality of distribution. A univariate, one-way repeated measures (RM) ANOVA will be conducted with factors group (live music intervention, recorded music active control, and standard care control) by time points (1, 2, 3, 4). The α significance threshold will be .05. The result of interest is the significance of the training×time point interaction, that is, whether the course of the data across time points was altered by the training type. Data are planned with 4 time points; in case of up to 2 missing time points, data will be interpolated [37]. Cases with more than 2 missing time points will be excluded from the analysis. Alternatively, if the residuals of the ANOVA are not normally distributed, a similar Kruskal-Wallis one-way ANOVA will be conducted.

For metrics that are collected on the same schedule as the VAS (EEG alpha power, delta/alpha power ratio, entropy metrics, HR, respiration rate, and blood pressure), the values per participant will be entered into the same F test described for the VAS ratings (as well as correlated with other baseline parameters as described in the “Other Parameters” section). For blood marker metrics that have 2 data collection points, a similar RM ANOVA will be conducted, but with 2 as opposed to 4 measures, and a significant interaction between treatment and time point would indicate whether treatment significantly impacted the amount of the neuroinflammatory blood marker. Categorical incidence metrics (delirium and POCD) will be evaluated with a one-way ANOVA to determine whether treatment impacted these neuroinflammatory-related complications. Delirium and signs of delirium will be perceived using the DOSS. The DOSS is a screening instrument for delirium, consisting of 13 items, and will be filled by a registered nurse during hospital admission until discharge. Nurses have frequent contact with the patients and are in a good position to observe behavior changes 24 hours a day when patients are admitted. The DOSS is already part of a routine in the surgical wards and all registered nurses are already trained. Likewise, scalar pain medication dosage will be averaged over 24-hour periods per patient and entered into a chi-square test to determine whether the dosage was significantly different across the 3 treatment groups. As this is exploratory and serves to inform future hypothesis-directed testing, we will not correct for multiple comparisons.

Posthoc analyses will be conducted to inform recommendations for participant grouping in future studies. Continuous baseline parameters (eg, age; see “Other Parameters” section for a full list) will be treated as covariates in an ANOVA in order to see whether the addition of these elements improves the model for each outcome. Similarly, categorical baseline parameters (eg, type of surgery) will be coded and entered as an additional factor (two-way RM ANOVA: treatment×baseline×time point) to see whether the additional factor can explain more variance than in the primary analysis model.

Previous studies have shown that postoperative pain management can be supplemented with music intervention. Yet, little is known about the benefits of live versus recorded music, or what physiological mechanisms may be mitigating pain perception as a response to listening to music. This NCT will compare the MiMiC live music intervention with listening to recorded music and care-as-usual. The primary objective is to see whether pain VAS ratings differ across the 3 conditions, and the secondary objective is to see whether postoperative neuroinflammation is reduced as a response to listening to music.

The strength of this study is the inclusion of preoperative assessments to establish some baseline characteristics of patients. First, the preoperative musical background questionnaire (selected items from [30]) assesses the estimated weekly hours of music listening as well as listening preferences for musical genre and whether or how the participant is musically trained. This provides us the opportunity to analyze how the extent of musical exposure may affect the primary outcome (pain VAS) of the intervention; depending on the strength of these findings, we may be able to even help inform future music interventions according to preoperative musical habits. Second, the preoperative cognitive assessment (6CIT) [28] serves as an important baseline for several secondary outcomes that may help to account for variability among the 3 intervention arms that otherwise might be attributed to the treatment. Specifically, the 6CIT will inform the interpretation of the postoperative cognitive status, and it may predict alpha power in EEG recording [22] that in turn may be linked to delirium [23,24]. If this proves to be the case, the preoperative cognitive status might be investigated in the future as a method to inform more at-risk patients who could especially benefit from supplementary nonpharmacological interventions [41] such as live or recorded music [42].

While postoperative pain is common to most patients undergoing surgical intervention in hospital, other characteristics surrounding the surgery (eg, type of surgery, duration of surgery, and underlying medical need for surgery) are widely varied across patients. Thus, a limitation of the study is that we will have a heterogeneous patient group. Considering that reduced pain VAS among patients experiencing MiMiC live music intervention compared to care-as-usual was previously a robust finding in a heterogeneous patient group [1], we do not consider our primary outcome to be affected by this heterogeneity. However, secondary outcomes related to finding neurophysiological markers of inflammation, and whether they are influenced by a music intervention, might be greatly affected by this heterogeneity. Moreover, the short duration of the intervention may further limit proper interpretation of inflammatory biomarkers, since previous studies have found neuroinflammatory biomarkers readings to be informative to overall interpretation 5 to 14 days after a medical event [43,44]. We may therefore form posthoc groups with similar surgery types when evaluating secondary outcomes, which could inform future hypotheses for further, longer term clinical research on music interventions and postsurgical neuroinflammation.

Upon completing this study, we will contribute a systematic comparison of the effects of live versus recorded music on postoperative pain. This study can ultimately inform policy makers as to the benefits to patient pain outcomes when incorporating live versus recorded music in postoperative recovery settings.