The Mere-measurement Effect of Patient-reported Outcomes

There is currently sufficient evidence to suggest that attributes such as the wording of questions, their presentation order, the context where they are asked, and the item's social sensitivity (e.g. I do not abuse my prescriptions), have an effect on participant responses (1,2). This family of variables are most oft referred to as method effects. Method effects largely regard the phenomenon in which the presentation of an item affects the participant's response independent of the content in question. The end goal within this area of study is to correct for any response bias resulting from the items used. This research primarily concerns psychometrics. It is important to note that while method effects relate to the research questions proposed, it is not the key feature of interest. The ensuing inquiry will not simply focus directly on changes to item responses, but instead, on changes to patient perceptions and behaviours as a result of their exposure to the measures themselves. This has been coined the mere-measurement effect (3) Research supports the claim that asking certain questions can influence behaviours on the same topic (4,5). For example, a randomized-control trial was conducted to assess the impact of asking questions pertaining to blood donation habits of the participants. The researchers found that the subjects whom were asked about their habits were significantly more likely to donate blood than their control counter parts at the following blood-drive (6). Other researchers found that when clinicians did not ask patients with Attention Deficit Hyperactivity Disorder (ADHD) to report whether they were currently medicated on a disease-related severity questionnaire, patients were more likely to report un-medicated symptomologies (7). In other words, patients on ADHD medications responded more symptomatically similar to their un-medicated selves when their medication status was not included on the questionnaire. This suggests that the medication status item itself, may have mitigated ADHD symptoms by bringing the patient's treatment into a higher state of awareness/directing attention.

Thus, it may be that asking patients about their intent to have an operation increases the likelihood of them opting for it, or, that asking them about their exercise increases their physical activity level. If so, this may have significant implications for the effect of routinely collecting patient-reported outcomes (PROs), as well as for utilizing measurements as interventions. However, before value can be attached to such an effect, first the presence and strength must be evaluated in the patient-specific context.

To date, no mere-measurement effect factors have been investigated in the context of patient-reported outcomes measures (PROMs) and particularly not in regard to how they may influence patient perceptions and behaviours. Therefore, exploration of this family of effects must be reduced to a few key anticipated features, namely, emotional valence (positive/negative) and frequency. For instance, would wording a cleft-lip question negatively, such as "I do not like my face", actually decrease the patient's preference for their own face, and, does repeated exposure to this question increase the magnitude of the effect.

This randomized controlled trial (RCT) review will be the first to the authors' knowledge to examine if measuring patient- reported outcomes directly effects their corresponding behaviour, and if this effect can be attributed in part to item wording and/or frequency of exposure.

PROMs \& Speech Disorders Three recent reviews regarding patient-reported outcomes in speech pathologies were identified (8,9,10). Francis et al. list a number of PRO measures that were designed for administration with subjects with any kind of voice disorders. The Evaluating Voice Disability-Quality of Life Questionaire (EVD-QOL) aims at assessing the general range of functional problems in voice disorders. The 30-items Voice Handicap Index (VHI) and its 10-items short version VHI-10 aim at providing a psychometrically robust voice diasability/handicap inventory. The Voice-Related Quality of Life (V-RQOL) aims at measuring the impact of voice problems on the quality of life. The Voice Activity and Participation Profile (VAPP) aims at identifying voice activity limitation and participation restriction separately. The Voice Symptom Scale (VoiSS) is an inventory of voice symptoms for assessing baseline pathology and response to change. The Communicative Participation Item Bank (CPIB) aims at assessing communication participation in all kinds of communication disorders. The Voice Self-Efficacy Questionnaire (VSEQ) monitors self-efficacy in individuals with self-declared voice problems before and after interventions. The Vocal Fatigue Handicap Questionnaire (VFHQ) and the Vocal Fatigue Index (VFI) aim at reflecting vocal fatigue.

Additional PROs include the Aging Voice Index, the Evaluation of the Ability to Sing Easily, the Glottal Function Index, the Linear Analog Scale of Assessment of Voice Quality, the Speech Disability Questionnaire, the Trans Woman Voice Questionnaire, the Vocal Cord Dysfunction Questionnaire, and the Vocal Tract Discomfort Scale Slavych, Zraick \& Ruleman, 2021.

The researchers have opted to use the CPIB and the VHI because these are well established and the former covers all kinds of communication disorders, while the VHI is using particularly negative wording in its original form.

1.2 STUDY RATIONALE The purpose of the mere-measurement study is to identify the most advantageous way to measure outcomes for patients. This will be accomplished by collecting measures on the same topic at different time intervals and with different wording of questions. The findings of this study should help researchers and clinicians collect measurements from patients in the least burdensome and most beneficial manner possible.

The primary objective of the mere-measurement study is to assess the effects of different collection methods on the responders wellbeing.

The secondary objectives of this study are to:

* What frequency of collection is best for patients? * What style of question wording is best for patients?

Research Questions:

* Does the assigned positivity or negativity of an item effect patient perception or behaviour on the topic inquired? * Does frequency of exposure to an item have an effect on patient perception or behaviour of the topic inquired?

Hypotheses:

H1 = The more a patient reads and responds to questions about their speech condition the more their speech condition will be affected.

H2 = The directionality of impact on a speech condition will be congruent with the emotional valence of the items used.

STUDY DESIGN Once a subject has agreed to participate in this study, they will be asked to complete online questionnaires as well as record their voice reading a collection of random words 2-4 times.

This study follows a 2x2 design. Participants will be randomly assigned to one of four groups. Randomization will be accomplished by using the MUW randomizer. Participants will be stratified by education and speech disorder type. The healthy case control group will undergo case matching to ensure comparability across groups. Each group will receive two brief speech pathology severity questionnaires but at different frequencies, and with slight adjustments to the wording. Depending on the experimental group, a subject will complete the questionnaires either two or four times, separated each time by one week. The questionnaires, in all groups, will take approximately 15 minutes to complete. Lastly, at the very start and once at the very end of the study, every participant will also be asked to record a brief 45 second audio file in the online survey as well as complete two additional psychological questionnaires.

The measures in this study cover the following tools (complete measures are uploaded in the ECS):

* A patient-reported outcome on voice health - Voice Handicap Index (VHI) or Voice Handicap Index Positively adapted (VHI-PA) * A patient-reported outcome on speech health - Communicative Participation Item Bank (CPIB) or Communicative Participation Item Bank Positively adapted (CPIB-PA) * A self-esteem stability scale - Self-Esteem Stability Scale (SESS) * A self-esteem scale - Rosenberg Self-Esteem Scale (RSS) * Two question measure of disease impact (2DB) * A single-question measure of disease activity (1DS) * Audio recording of text read aloud (ARec)

Study population All patients with Muscle tension dysphonia, Inducible laryngeal obstruction, Amyotrophic lateral sclerosis (ALS), suffering from a stroke or other brain injury/damage/trauma, (aphasia, dysarthria), and Parkinson's disease will be targeted for recruitment. There are no geographic location requirements. The participation schedule can be found below. Patients will be informed of the schedule in advance. A subject must complete any respective sampling epoch within 48 hours of the prescribed date in order to qualify as completed. Subjects must also have strong English skills and be technologically savvy to participate. Only adults will be included in the study.

Study Design Flow Group 1: High frequency-negative: 1 week between each collection

1. Demographics, VHI, CPIB, SESS, RSS, 1DS, 1DB, ARec\* 2. VHI, CPIB, SE, SSE, 1DS, 2DB, ARec 3. VHI, CPIB, SE, SSE, 1DS, 2DB, ARec 4. VHI, CPIB, SE, SSE, 1DS, 2DB, ARec Group 2: High frequency-positive: 1 week between each collection

<!-- -->

1. Demographics, VHI-PA, CPIB-PA, SESS, RSS, 1DS, 2DB, ARec\* 2. VHI-PA, CPIB-PA, SE, SSE, 1DS, 2DB, ARec 3. VHI-PA, CPIB-PA, SE, SSE, 1DS, 2DB, ARec 4. VHI-PA, CPIB-PA, SE, SSE, 1DS, 2DB, ARec Group 3: Low frequency-negative: 2 weeks between each collection

1: Demographics, VHI, CPIB, SESS, RSS, 1DS, 2DB, ARec\* 2: VHI, CPIB, SE, SSE, 1DS, 2DB, ARec Group 4: Low frequency-positive: 2 weeks between each collection

1. Demographics, VHI-PA, CPIB-PA, SESS, RSS, 1DS, 2DB, 1ARec\* 2. VHI-PA, CPIB-PA, SE, SSE, 1DS, 2DB, 1ARec Control: (once at start, once at low frequency completion/final outcome collection, once at high frequency completion)

1: Demographics, ARec x 3

Healthy Control: (once at start, once at low frequency completion/ final outcome collection, once at high frequency completion)

1. Demographics, ARec x 3

\*completed prior to assessments

\*English was chosen as the study language as a wider range of validated assessments exist in this language. The AI used for speech analysis is also best suited for and trained (66%) on English. Lastly, English is the most prolific language (native + secondary speakers) in the world which will facilitate the speed of recruitment.

Subjects who meet any of the following criteria will be excluded from study entry:

• Under the age of 18

NUMBER OF SUBJECTS PARTICIPATING IN THE STUDY Subjects will be randomly separated into six groups including control and healthy control, with approximately 33 subjects per group. A formal power analysis was performed using G\*power with parameters sourced from relevant seminal research (Conner et al., 2011; Godin et al., 2011). The following parameters were used for the sample size analysis: two-tailed directionality; effect size of 0.30; a priori computation, anticipated power of 0.95, and an error rate of 0.0023. Means difference tests between independent groups within the t-test family were utilized.

RECRUITMENT PROCEDURE Recruitment into the mere-measurement study will be performed supported by Probando. Probando is a previously approved patient recruitment service that specializes in online advertising. All recruitment materials are uploaded to the ECS. A description of the Probando recruitment service and procedures can be found as an attachment, as well as the recruitment materials for all subjects including healthy controls and a screenshot of the currently inactive landing page. As a part o

• Self-Esteem Stability Scale (SESS) [Time frame: Baseline and at each experimental instance occuring once a week for up to four weeks]
• Disease activity (1DS) [Time frame: Baseline and at each experimental instance occuring once a week for up to four weeks]
• A self-esteem scale - Rosenberg Self-Esteem Scale (RSS) [Time frame: Baseline and at each experimental instance occuring once a week for up to four weeks]
• Disease Impact [Time frame: Baseline and at each experimental instance occuring once a week for up to four weeks]