Fatigue, Anxiety, Music, and Entertainment (FAME) Study (FAME)

Addressing Fatigue, Anxiety, and Cognitive Impairment Through Rhythmic Effects: A Pilot Feasibility Study Using a Group Drumming Intervention With Cancer Patients

Evaluate the feasibility of a six-week group drumming intervention on fatigue, anxiety, and cognitive impairment when compared to an attentional control for cancer patients who have undergone at least one treatment session of chemotherapy or radiation therapy.

The AIMS related to specific outcomes are: AIM 1. Evaluate the feasibility of a group drumming intervention. RQ 1.1 What are the rates of recruitment? RQ 1.2 What are the potential barriers to recruitment? RQ 1.3 Was the group drumming intervention protocol delivered as designed? RQ 1.4 Did the attention control group follow designed protocol? RQ 1.5 What are the differences in treatment acceptability? AIM 2. Evaluate adherence and duration between a group drumming intervention and attention control group. RQ 2.1 What are the adherence rates for the 6-week interventions among cancer patients, and were these rates similar in the two treatment arms? RQ 2.2 What are the dropout rates for the 6-week interventions among cancer patients, and were these rates similar in the two treatment arms? RQ 2.3 Are there differential reasons for drop out based on group assignment? AIM 3. To obtain preliminary data describing the impact of group drumming on fatigue and anxiety severity levels and change patterns over time as compared to an attentional control group. Exploratory AIMs. To explore whether cancer patients, who participate in a 6-week group drumming intervention, will show improvement in cognitive function, as measured by the Functional Assessment of Cancer Therapy-Cognition (FACT-COG). To explore impact on group drumming and positive affect as measured by Positive Affect & Well-being Scale-Short Form (PAW-SF) and Core Flow States (C FSS). Statistical Methods, Data Analysis, and Interpretation. Demographics, feasibility measures, and outcome measures will be analyzed through a combination of descriptive and inferential analysis. Sample characteristics will be reported as means and quartiles, counts, or percentages, as appropriate, along with standard deviations or interquartile ranges to describe variability. Where it is possible to construe sample values as population estimates, standard errors and confidence intervals will also be provided. Comparable information will be provided for the samples randomly assigned to receive either intervention or attention control. Any chance imbalance will be controlled statistically by conditioning on baseline values, as described in many of the statistical references. Drop out and attendance percentages will be calculated within each group and evaluated as potential treatment responses. Aim 1: This AIM focuses on determining the pragmatic possibility of success in recruiting and delivering a group drumming intervention. Rates of recruitment will include reporting the percentage of people who expressed an interest in participating from those approached, percentage of people who met inclusion criteria from those who expressed an interest, percentage of people consented from those who met inclusion criteria, the percentage of people consented from those who met inclusion criteria, the percentage of people who completed the questionnaires. Barriers to recruitment will include reporting percentage of those who were excluded from those who expressed an interest, documentation of barriers people stated for not being able to participate in the study with percentages for each reason. A CONSORT flowchart will be used to show how many people were screened, how many were eligible or excluded, how many consented, and how many completed the study. Fidelity related to following protocols for the group drumming intervention will be evaluated by using a music therapy validity checklist. The music therapist facilitating the intervention will sign a copy of the protocol and fidelity checklist prior to beginning the intervention and the percentage of deviations will be recorded on the protocol fidelity check list by the principal investigator during the first intervention group to be sure that the protocol is being followed in order to ensure fidelity. Aim 2: Adherence (compliance, persistence, "dose") between a group drumming intervention and attention control group will be calculated based upon percentages of participants who missed a group, percentage of participants who dropped out, percentage of people who did not complete the study, and percentage of people who completed the study. Confidence intervals for the percentages will be derived from exact permutational tests that require no distributional assumptions and are hence more appropriate for small samples than those based on normal theory. Duration will be examined using discrete time survival analysis and bootstrapped confidence intervals. Reasons for missing a group will be tallied (see appendix 9), and grouped into "unrelated" or "possibly treatment-related" categories. Examples of "presumptively unrelated" reasons would be colds or minor illness, dental or medical appointments, family issues, or moved to a different area. If participants missed a group session because they could not accept or tolerate the experience, this would be coded as possibly treatment-related. Those who actively withdraw (letting music therapist or research team member know ahead of group or after group that they don't plan to return) will be noted separately as neutral reasons from those who passively withdraw (don't show up). Similarly, the percentage of positive and negative answers to "How did the session go today?" will be reported. The answers to the two questions "what made this study challenging for you to attend?" and "what didn't you like about it?" will be reported and categorized into percentages. Exploratory Aim 3: This aim provides preliminary pilot data describing the impact of group drumming on fatigue, anxiety, and cognitive impairment severity levels and change patterns. These data will be needed to plan sample size and design features in a future definitive comparative study. For this purpose, a formal statistical model is most useful, since it generates parameter estimates that can be applied to power calculations for similar model-based analyses to be used in the future study. Two aspects of an individual's response are relevant: the degree of overall treatment impact (gain) and the mean rate-of-change post-baseline (trend). Fatigue will be analyzed by looking at gain and trend of individual FACIT-F scores at 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups. Anxiety will be analyzed by looking at gain and trend of individual using STAI scores at 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups. Cognitive impairment will be analyzed by looking at gain and trend of individual using FACT-COG scores at 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups. Analysis model. The two primary outcomes of fatigue and anxiety will be analyzed using a mixed effects linear model that assumes that individual longitudinal responses, conditional on baseline values, are generated by treatment assignment, and time since baseline. Specifically, the post-baseline measures at three, six, and eight weeks after the last group ends will be analyzed by examining the difference in mean adjusted gains and trends. Equally important in this kind of pilot study, the model yields empirical Bayes estimates of individual gains and time trends. Such a clinical impact analysis model will allow the study to show individual responses related to who is benefitting, how much they are benefitting so that responses can be viewed based on trend and adjusted gain (or loss), and how varied these responses are relative to the mean treatment impact. Details of the model are developed in multiple scientific publications. The following hypotheses will be explored under the clinical impact model: HImpact: Drumming yields greater mean improvement on fatigue, anxiety and cognitive impairment than does attention control. Hpersistence: The mean rate-of-change with drumming is more positive than that of the control treatment, Hheterogeneity: Patients receiving either intervention vary in the extent of their benefit. Data will be analyzed using SPSS Mixed and SAS Proc Mixed programs under inferential options appropriate for smaller samples (Restricted Maximum Likelihood with Kenward-Roger Degrees of Freedom). The hypothesis represents plausible expectations, not formal statistical assumptions of population magnitudes. The purpose of the analyses is to generate estimates for future planning, not to reach definitive conclusions. I plan to recruit 26 participants (n=13 treatment, n=13 control) allowing for 20% drop out to yield 10 per group will allow me to plot the distributions of the individual and mean estimates in a scatterplot with adjusted gain and trend as the two axes.

Utilizing a randomized group pilot research design, the feasibility and exploratory impact of a group drumming intervention on fatigue, anxiety, and cognitive impairment in cancer patients will be compared to an attention control. Participants will be randomized to a six-week group drumming intervention or a six-week attention control group. The attention group is intended to control for time effects and the nonspecific effects of participating in a group learning activity.

6 week one hour/week drumming sessions. The detailed intervention protocol was created in collaboration with board-certified music therapists. The music used in the intervention will include both recorded percussion music and the use of percussion instruments.

6 week one hour/week educational series. The educational group is largely intended to create an attention group control so group effect and time of the drumming intervention can be controlled.

Cancer patients will be led by a Board Certified Music Therapist to participate in group drumming for one hour per week, for six weeks.

Fatigue will be analyzed by looking at the degree of overall treatment impact (gain) changes over time in group and individual Functional Assessment of Chronic Illness Therapy Fatigue (FACIT-F) scores.

FACIT-F scores can range between 0 and 52 (0 is the worst possible score). The FACIT-F is a 13-item symptom subscale measuring fatigue with chronic illness that has been validated among various cancer diagnosis and has been used in interventional, repeat measure studies with a test-retest reliability of r=0.90 and internal consistency of a=0.95 (Minton & Stone, 2008).

Baseline, 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups.

The mean rate-of-change patterns (trend) of group drumming on fatigue over time as compared to an attention control group will be measured using FACIT-F

The mean rate-of-change for the group post-baseline (trend) will be analyzed. Fatigue will also be analyzed by looking at trend of individual FACIT-F scores that can range between 0 and 52. The FACIT-F is a 13-item symptom subscale measuring fatigue with chronic illness that has been validated among various cancer diagnosis and has been used in interventional, repeat measure studies with a test-retest reliability of r=0.90 and internal consistency of a=0.95 (Minton & Stone, 2008).

Baseline, 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups.

Anxiety severity levels of change over time will be compared to an attentional control group using the STAI-Y form.

Anxiety will be analyzed by looking at gain of individual State section of State Trait Anxiety Inventory-State Scale (STAI) scores at 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups. The STAI-form Y-1 will be used to measure state anxiety. It has been used with cancer patients to measure state anxiety with a Cronbach's alpha score of .94 for state anxiety. STAI-form Y-1-inventory consists of 20 anxiety questions like "I feel calm" with participants rating their answers on a four-point Likert scale (1=not at all, 4=very much so). STAI-Y has a minimum score of 20 to a maximum score of 80. STAI scores are commonly classified as "no or low anxiety" (20-37), "moderate anxiety" (38-44), and "high anxiety" (45-80).

Baseline, 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups.

Anxiety change patterns over time will be compared to an attentional control group using the STAI-Y form.

Anxiety will be analyzed by looking at the trend of individual using State section of State Trait Anxiety Inventory-State Scale (STAI) scores at 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups. The STAI-form Y-1 will be used to measure state anxiety. It has been used with cancer patients to measure state anxiety with a Cronbach's alpha score of .94 for state anxiety. STAI-form Y-1-inventory consists of 20 anxiety questions like "I feel calm" with participants rating their answers on a four-point Likert scale (1=not at all, 4=very much so). STAI scores are commonly classified as "no or low anxiety" (20-37), "moderate anxiety" (38-44), and "high anxiety" (45-80).

Baseline, 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups.

Cognitive function of cancer patients who participate in a 6-week group drumming intervention will show improvement in cognitive function, as measured by the Functional Assessment of Cancer Therapy-Cognition (FACT-COG).

Cognitive impairment will be analyzed by looking at gain and trend of individual using FACT-COG scores. The FACT-Cog contains 37 items, with subscales consisting of 1) patients' perceived cognitive impairments, 2) perceived cognitive abilities, 3) noticeability or comments from others, and 4) impact of cognitive changes on quality of life. Scores can range from 0 to 148 and a higher score indicates better perceived cognitive function.

Baseline, 3 weeks, 6 weeks, and 8 weeks after group has ended between intervention and attention control groups.

Positive affect of group drumming will be measured over time using the Positive Affect & Well-being Scale-Short Form (PAW-SF).

Measured by Positive Affect & Well-being Scale-Short Form (PAW-SF). The 9-item positive affect and well-being short form (PAW-SF) was developed to measure positive affect, life satisfaction, and an overall sense of purpose and meaning and responds to longitudinal changes as a result of a treatment with test-retest reliability of 7 days (Salsman et al., 2013). The test-retest correlation coefficients range from 0.59 to 0.86 and the Cronbach's alphas ranges from 0.94 to .095. Raw scores range from 9 to 45, T-scores range from 26.3 to 68, and Standard Errors range from 3.9 to 4.5 (the higher scores result in a better outcome).

The Core Flow States Scale (C FSS) was based upon the work of the Mihaly Csikszentmihalyi who developed the concept of flow to describe the sense of being totally absorbed in a way that feels positive and the scale has been determined to be a valid and reliable measure to capture one's subjective state of being in flow. Internal consistency estimates ranged from .91 to .94. There are 10 questions with five choices ranging from 1 = strongly disagree to 5=strongly agree. Summed scores range from 10 to 50 and each individual sum is divided by 10 resulting in a final score of 1 being the lowest and 5 being the best score to measure a sense of flow.

Inclusion Criteria: Adults 18 and over who have received at least one session of chemotherapy or radiation treatment and self-report symptoms of fatigue or anxiety at a level 4 or above on a scale of 0-10. Active treatment through one-year post completion of active treatment. At least one week after surgery. Ability to read and write in English. Exclusion Criteria: Total deafness or severely impaired hearing. Absence of arms and/or hands. Current engagement in group drumming. Scheduled for surgery in the next 12 weeks. Currently taking medications for diagnosis of dementia or Alzheimer's disease. Substantial, uncorrected vision loss.

Fairclough, D.L., Design and analysis of quality of life studies in clinical trials. 2010: CRC press.

Moinpour CM, Darke AK, Donaldson GW, Thompson IM Jr, Langley C, Ankerst DP, Patrick DL, Ware JE Jr, Ganz PA, Shumaker SA, Lippman SM, Coltman CA Jr. Longitudinal analysis of sexual function reported by men in the Prostate Cancer Prevention Trial. J Natl Cancer Inst. 2007 Jul 4;99(13):1025-35. doi: 10.1093/jnci/djm023. Epub 2007 Jun 27.

Donaldson GW, Moinpour CM. Individual differences in quality-of-life treatment response. Med Care. 2002 Jun;40(6 Suppl):III39-53. doi: 10.1097/00005650-200206001-00007.

Moinpour CM, Lovato LC, Thompson IM Jr, Ware JE Jr, Ganz PA, Patrick DL, Shumaker SA, Donaldson GW, Ryan A, Coltman CA Jr. Profile of men randomized to the prostate cancer prevention trial: baseline health-related quality of life, urinary and sexual functioning, and health behaviors. J Clin Oncol. 2000 May;18(9):1942-53. doi: 10.1200/JCO.2000.18.9.1942.

Donaldson GW, Moinpour CM. Learning to live with missing quality-of-life data in advanced-stage disease trials. J Clin Oncol. 2005 Oct 20;23(30):7380-4. doi: 10.1200/JCO.2005.07.022. Epub 2005 Sep 26. No abstract available.

Moinpour CM, Donaldson GW, Redman MW. Do general dimensions of quality of life add clinical value to symptom data? J Natl Cancer Inst Monogr. 2007;(37):31-8. doi: 10.1093/jncimonographs/lgm007.

Moinpour CM, Donaldson GW, Liepa AM, Melemed AS, O'Shaughnessy J, Albain KS. Evaluating health-related quality-of-life therapeutic effectiveness in a clinical trial with extensive nonignorable missing data and heterogeneous response: results from a phase III randomized trial of gemcitabine plus paclitaxel versus paclitaxel monotherapy in patients with metastatic breast cancer. Qual Life Res. 2012 Jun;21(5):765-75. doi: 10.1007/s11136-011-9999-z. Epub 2011 Sep 16.

Moinpour CM, Darke AK, Donaldson GW, Cespedes D, Johnson CR, Ganz PA, Patrick DL, Ware JE Jr, Shumaker SA, Meyskens FL, Thompson IM Jr. Health-related quality-of-life findings for the prostate cancer prevention trial. J Natl Cancer Inst. 2012 Sep 19;104(18):1373-85. doi: 10.1093/jnci/djs359. Epub 2012 Sep 12.

Donaldson GW, Nakamura Y, Moinpour C. Mediators, moderators, and modulators of causal effects in clinical trials--Dynamically Modified Outcomes (DYNAMO) in health-related quality of life. Qual Life Res. 2009 Mar;18(2):137-45. doi: 10.1007/s11136-008-9439-x. Epub 2009 Jan 20.

Moinpour CM, Sawyers Triplett J, McKnight B, Lovato LC, Upchurch C, Leichman CG, Muggia FM, Tanaka L, James WA, Lennard M, Meyskens FL Jr. Challenges posed by non-random missing quality of life data in an advanced-stage colorectal cancer clinical trial. Psychooncology. 2000 Jul-Aug;9(4):340-54. doi: 10.1002/1099-1611(200007/08)9:43.0.co;2-f.

Coffman CJ, Edelman D, Woolson RF. To condition or not condition? Analysing 'change' in longitudinal randomised controlled trials. BMJ Open. 2016 Dec 30;6(12):e013096. doi: 10.1136/bmjopen-2016-013096.

Liu GF, Lu K, Mogg R, Mallick M, Mehrotra DV. Should baseline be a covariate or dependent variable in analyses of change from baseline in clinical trials? Stat Med. 2009 Sep 10;28(20):2509-30. doi: 10.1002/sim.3639.

Liang, K.-Y. and S.L. Zeger, Longitudinal data analysis of continuous and discrete responses for pre-post designs. Sankhyā: The Indian Journal of Statistics, Series B, 2000: p. 134-148.

Singer, J.D., J.B. Willett, and J.B. Willett, Applied longitudinal data analysis: Modeling change and event occurrence. 2003: Oxford university press.