Table 3: Evidence table.
|
Author/Year |
Level of Evidence
Study Design
Risk of Bias |
Participants
Inclusion Criteria
Study Setting |
Intervention
Control Groups |
Outcome Measures |
Results |
Strengths
Weaknesses |
|
Betsch, et al. [30] |
Level 3B
Cross Sectional Cohort Study
Risk of Bias Low |
Participants N = 50 participants for standing measures (24F/26M) (Mean age = 25.3) 34 Participants for dynamic measures (16F/18M) (Mean age = 25.4)
Inclusion Criteria - Volunteers from the university community - Aged 18-50-years-old - more than one year experience using a touch smartphone - daily smartphone usage > 60 minutes.
Study Setting Simulation in a laboratory |
Intervention - Each participant conducted a one-handed and two-handed texting task while standing and while walking on a treadmill. - Participants simulated a phone call with the smartphone to their ear while standing and walking on a treadmill. - As a control, participants were standing in a neutral standing position, and while walking on a treadmill without a smartphone. - Participants completed the smartphone addiction scale and SF-36 health questionnaire. |
Outcome Measures - Spinal posture and pelvic position during standing and while walking on a treadmill was measured using a surface topography system.
- Self-report smartphone addiction scale (SAS)
Devices Tested - Smartphone |
Significant Findings - All smartphone tasks lead to a significant increase in thoracic kyphosis and trunk inclination during standing and while walking. - A significantly increased lumbar lordosis was also found. - Texting with one or two hands correlated with increased surface rotation. - No associations between smartphone addiction and changes in spinal posture were reported. |
Strengths - Reliable system used to measure spinal posture and pelvis position. - Extensive statistical analysis. - Control measurements recorded prior to testing.
Weaknesses - Study design does not allow us to draw conclusions about changes in spinal posture over time with smartphone use. - Size of sample. - Some participants did not want to undergo dynamic measurements as it meant revealing their trunk. |
|
Choi, et al. [31] |
Level 3B
Cohort Study
Risk of Bias Low |
Participants 20 subjects (11M/9F) (Mean age = 21.5)
Inclusion Criteria - No previous or current musculoskeletal disorders. - Have owned a touch-smartphone for more than three years. - No difficulty in using a smartphone while walking.
Study Setting Simulation in a laboratory |
Intervention Walking on a treadmill in five different conditions: - normal walking without using a phone, - conducting one-handed browsing while walking, - two-handed texting while walking, - walking with one arm bound, - walking with both arms bound.
|
Outcome Measurements - Spine kinematics variables and the myoelectric activity levels of the lumbar erector spinae muscles were quantified and compared between the five walking conditions
Devices Tested - Smartphone |
Significant Findings - Thoracic kyphosis and lumbar lordosis was more significant when using a phone (P < 0.05). - The median level of muscle activity was 16.5% (browsing) to 31.8% (texting) greater than walking without a smartphone, with the differences being significant (P < 0.05). - No significant difference in muscle activity found between the two bound walking conditions. |
Strengths - Data collected and quantified by reliable systems. - A reference posture was recorded before the experiment for each participant.
Weaknesses - Examination of muscle activity and spine kinematics observed only for 5 minutes. - Simulation of walking in a laboratory may not represent real usage environments of smartphone use on the street. - Potential effects of environmental factors were not considered. |
|
Douglas & Gallagher [22] |
Level 2B
Prospective cohort study
Risk of Bias Low |
Participants N= 22 young adults (11M/11F) 20-23 yrs old.
Inclusion Criteria - No previous neck or spine injury - No chronic headaches - No allergies to rubbing alcohol - No exposure to lumbar spine x-ray, GI tract x-ray, barium enema x-ray, or CT scans within the past two years. - Women- no chance of pregnancy
Study Setting University of Arkansas- Walker Health Center |
Intervention - X-rays taken while participants were seated in 5 different postures (neutral, max neck flexion, upright seated, semi-reclined, reclined) with biomarkers placed on 4 different part of their body |
Relevant Measures taken - Participant were instructed to read/look at a tablet without using the arm rests - Measures taken were centroid cervical lordosis (CCL), skull angle relative to the horizontal, gravitational moment arm (GMA), And intervertebral Joint angles between skull-C1, C1 -C2, C2-C3.
Devices Tested -Tablet |
SignificantFindings - CCL was more flexed in each position compared to neutral. - There was a significant effect of trunk position for the moment arm. GMA of the head was smallest when reading a tablet with a recline trunk vs. semi-reclined and upright. Upright position was not significantly different. - Trunk position also had a significant effect on skull angle. Skull angle was less flexed in neutral compared to the 3 reading postures. Of the 3 reading postures, there was less skull flexion in the reclined position vs. semi and upright; however there was no difference between semi vs. upright. - There was a significant main effect of trunk position for skull-C1, C1-C2, and C2-C3. Skull-C-1 Full flexion was significantly more flexed vs. other four positions. C1-C2 was more extended in neutral and reclined vs. upright position, but was not different compared to semi reclined and upright positions. C2-C3 was more flexed in reclined compared to neutral. |
Strengths - Multiple (5) angles were measured.
Weaknesses - The study only examined a short duration, simulated reading tablet task. - Hip and trunk angle were not specifically measured within the seat the participant was in other than cueing them to sit as far back as possible. Hip angle was not directly controlled for and could have been responsible for some variations in intervertebral joint angles. - The study did not assess thoracic angle or intervertebral joints below C7, and upper thoracic kyphosis could be a further risk of semi-recline or reclined positions. - Small sample size and were all university students. |
|
Elserty, et al. [23] |
Level 2B
Cross-sectional survey study
Risk of Bias Low |
Participants N = 420 Physical Therapy students (133M/280F) mean age of 18-21 yrs
Inclusion Criteria - Having smartphone to access survey - Being able to ready and understand survey in English - Being an undergraduate university student of physical therapy. - Using a smartphone for 2+hr on a typical day.
Study Setting - Egypt - Online |
Intervention - Online questionnaire administered at one point in time. |
Demographic data - Prevalence of smartphone use
Relevant measures taken - Different sites of discomfort (shoulder, arm, neck, eye, back, hand/finger, wrist, leg/feet. -Types of discomfort (stiffness, tingling/numbness, aching/pain, cramping, soreness) - Common positions during usage of smartphone (standing, sitting on floor, sitting in chair/couch, laying on side/side lying, lying on stomach/prone, lying on back/supine)
Devices Tested - Smartphone |
Significant Findings - 62.4% of participants reported smartphone addiction, which what higher in females (68%) than males (52%) - There was a significant difference between males and females feeling pain in wrist and back and neck. - The most common discomfort was felt at the neck region. - There was a significant difference between males and females regarding pain in shoulders, hands, and fingers. - There was a significance between duration of smartphone use over 6hrs per day and feeling discomfort with a specific body part. |
Strengths - Several body positions were included - Several locations for discomfort were included with the different types of discomfort
Weaknesses - Only one point in time was measured - The study was limited to only physical therapy students and not the general public - The sample size was small and there was no follow up questionnaire |
|
Emodi-Perlman, et al. [32] |
Level 3B
Cohort Study
Risk of Bias Moderate |
Participants 3 groups of young aged adults (18-35 years-old). Group 1 included ultra-Orthodox subjects. Group 2 included Orthodox subjects. Group 3 included secular subjects
Inclusion Criteria - Young adults - Good health, - Using a phone for remote communication.
Study Setting Questionnaire completed in person at participants schooling institution or place of work |
Intervention - Questionnaires on demographic variables, mobile phone use characteristics, anxiety, depression, daytime sleepiness, bruxism, and diagnosis of temporomandibular disorders |
Outcome Measures Self-Report Questionnaire
Device Tested Smartphones |
Significant Findings - Various aspects of smartphone use, including being awakened at night, stress caused by information delivered by phone, and stress from phone overuse increased the risk of daytime sleepiness, temporomandibular disorders, and bruxism.
|
Strengths - Questionnaire was filled out on the spot so any questions could be asked while completing the questionnaire
Weaknesses - Participants are not necessarily representative samples of their sectors. Several additional demographic, cultural, and lifestyle differences might have affected the results |
|
Gustafsson, et al. [24] |
Level 2B
Longitudinal Cohort Study
Risk of Bias Low |
Participants N = 7092 Young Adults (2759M/4333F) aged 20-24 years old
Inclusion Criteria - Responds to at least one question about SMS texting
Study Setting - Sweden - Web-based questionnaire |
Intervention - Survey administered at baseline, one year, and five years |
Text Messaging - Number of texts sent in last 30 days
Musculoskeletal Symptoms - Pain in upper back/neck - Pain in shoulders, arms, wrists, hands - Numbness/tingling in hands and fingers |
Significant Findings - For those with no symptoms at baseline, prospective associations were found between high reports of text messaging and numbness and tingling in the hands at one year follow-up. -For those with symptoms at baseline, associations were found between text messaging and reported pain in neck/upper back and in shoulder/upper extremities.
Nonsignificant Findings - At five year follow-up, reports were of pain in the shoulder/upper extremities, but no association was found between text-messaging and reported symptoms. |
Strengths - Large study group, population-based, longitudinal design, control for confounding variables.
Weaknesses - Self-report data tends toward potential bias, limited information on length of text messages or use of phone for other purposes, limited information on the nature of work or study occupations, only data from 3 points in time. |
|
Korpinen & Pääkkönen [33] |
Level 3B
Cohort Study
Risk of Bias Low |
Participants N = 1563 Young Adults (604M/956W) (30-years-old and younger) (Mean = 24.1 +/- 3.6 years)
Inclusion Criteria - Under 30-years-old
Study Setting - Finland - Web-based questionnaire |
Intervention - Self-Report questionnaire |
Use of Technology - Type of electronics used (Desktop Computers, Portable Computers & Mobile Phones) - Amount of use
Musculoskeletal Symptoms - Aches, pains, or numbness in various body segments |
Significant Findings - 65% of women experienced pain, numbness, or aches in the neck pretty often, compared to 34.5% of males. - Overall, 53.3% of young adults experienced symptoms in the neck and 32.2% in the hip and lower back. - Associations were found between use of desktop versus portable computers at leisure and symptoms in various body segments. -Exhaustion also had an association with some symptoms. |
Strengths - High response rate.
Weaknesses - Inability to differentiate between "home work" and "students" - Participants may have different interpretations of the symptoms. - Self-report tends towards a response bias. - Confounding variables may contribute to symptoms, such as stress. |
|
Lin, et al. [25] |
Level 2B
Cohort Study: 3x3 mixed factorial design
Risk of Bias Low |
Participants N = 18 adults - (9M/9F) - (Age 20-31 years)
Inclusion Criteria - Right hand dominant - No known musculoskeletal disorders in neck, back, buttocks, or extremities - No non-specific pain in areas stated above within past 6 months - Can touch type on electronic keyboard at least 35 words/min - No lifting of objects over 5 kg in week before study
Study Setting - Simulated setting in laboratory |
Intervention - Random assignment to desk, lap, or bed for usage position - 60 minute typing on Apple ipad, with keyboard being manipulated (STD, Wide, and Split) over 6 non-consecutive days
Control - Although participants were assigned randomly to 3 different conditions, there was no control group tested for reference |
Joint Position - Wrist, elbow, and neck joint angles measured by flexible electro-goniometry system - Wrist extension, flexion, and radial and ulnar deviation also measured by bi-axial electrogoniometers
Musculoskeletal Discomfort - 0-10 body discomfort scale
Electronic Use - System-usability scale questionnaire
Devices Used - Apple i-pad (tablet) |
Significant Findings - Difference in joint angles of upper limb and neck with different positions and type of keyboard used - Range of joint movement significantly increased with electronic usage time - Position of tablet and keyboard design affecting wrist extension (with bed being most, lap being next, and then table) - Self-perceived discomfort in UE increased with electronic use duration
Non-significant Findings - Lap position had highest self-perceived discomfort, and desk had lowest - Tablet use in bed showed more wrist extension but a more neutral elbow compared to tablet use at a desk. - The keyboard option with the least ulnar deviation was the angled split keyboard. - The wrists, elbows, and neck showed a 13-38% increase in joint movement from the middle of the typing session compared to the end. |
Strengths - Random assignment to testing groups
Weaknesses - Did not look at non-typing tasks - Posture not observed in a natural setting during this simulated task -Objective task performance was not analyzed - Data collected from myoelectrical signals from active muscles during task not analyzed - Contact pressure tablet can cause on users was not evaluated |
|
Pais, et al. [34] |
Level 3B
Cross-sectional cohort
Risk of Bias Low |
Participants - 88 participants total. 42 males and 46 females. Students and staff of a university. - Ages 18-25
Inclusion Criteria - Participants use digital devices 6 + months for a minimum 5 days/week - No previous trauma to cervical spine or any deformities
Study Setting - Not specified. Performed with university students and staff. |
Intervention - 4 groups separated individuals based on duration of usage of devices (1hr/day; 1-2 hr/day; 2-4 hr/day; 4+ hr/day)
Control - No reported control groups |
Outcome Measures - Measurements taken for cervical ROM. ROM tested 3 times with 1 minute intervals and average was recorded. - Images also taken for measuring craniovertebral angle
Devices used - Mobile phones |
Significant Findings - Reported statistical significance in flexion variable between the four groups - No difference noted in extension variables - Additionally, no difference in the CVA variable. |
Strengths - No reported conflict of interest; statistically significant differences
Weaknesses - Small sample size; participants dominantly used smartphones; no educational interventions to observe outcome variables |
|
Queiroz, et al. [35]
|
Level 3B
Cross-Sectional Cohort Study
Risk of Bias Low |
Participants N = 299 adolescents (10-19 years old)
Inclusion Criteria - No musculoskeletal pain secondary to infections, rheumatic, oncologic, genetics, DM, or thyroid diseases - No reports of recent trauma - Must have informed consent from legal guardian
Study Setting - Classroom within private school in São Paulo, Brazil |
Intervention - Self-report questionnaire reporting demographic data, physical activities, musculoskeletal pain symptoms, and the use of television/digital media (including computer, internet, electronic games, and cell phones). |
7 Musculoskeletal pain syndromes were evaluated: - Juvenile fibromyalgia - Benign joint hypermobility syndrome - Myofascial syndrome - Tendinitis - Bursitis - Epicondylitis - Complex regional pain syndrome
Devices used: - Computer - Mobile handheld devices - Nintendo DS - Gameboy - Wii - Playstation - Xbox |
Findings - 61% reported musculoskeletal pain with 66% being in the back, 49% in the neck, 41% in the lower limbs, 31% in the shoulder/arms, and 19% in the wrist and hands. - Electronic game use was reported by 70% of adolescents. - At least 1 musculoskeletal pain syndrome was observed in 33% of the adolescents. - Overall, the study presented a high frequency of musculoskeletal pain in adolescents, most commonly 15 year olds who use at least 2 electronic devices. |
Strengths - The use of a questionnaire with excellent test–retest reliability...reducing the effect of memory bias (change to Excellent test-retest reliability shown by Kappa index of 0.83, which reduced effect of memory bias) - Utilization of a systematic musculoskeletal physical examination of adolescents who complained of musculoskeletal pain, using established criteria for musculoskeletal pain syndromes.
Weaknesses - The population only includes upper and upper/middle socio-economic classes with low frequency of work activities - meaning work may increase pain tolerance from low-demand use. - Other issues were not studied such as emotional disorders and other risk behaviors - Self-report of electronic use presents potential bias |
|
Short, et al. [36]
|
Level 3B
Cross-Sectional Cohort Study
Risk of Bias Low
|
Participants N = 46 Adults - (44F/2M) - (Age 20-40 years)
Inclusion Criteria - Use of smartphone - No known musculoskeletal conditions impacting posture - Graduate students in doctor of occupational therapy (OTD) program at Huntington University
Study Setting - Laboratory setting at Huntington University
|
Intervention - Smart phone used to either search the internet or text while seated in a chair.
Control - No control group, with all participants experiencing the same condition of being seated in a chair and performing the same task.
|
Joint Position and Posture - Goniometer measurements including joint position and degree of motion of the digits, scapula, spine, shoulder, elbow, forearm, and wrist
Self Report Questionnaire and Apple Reported “Screen Time” - Data collection on how frequently phone is used (hours/day)
Devices Used: - Smartphone
|
Findings - Screen time showed an average of 143 minutes per day within mobile phone use - Results showed cervical spine flexion, scapular protraction, elbow flexion, and wrist ulnar deviation with thumb flexion during smart phone use that could contribute to deformity over time.
|
Strengths - Extensive goniometric measurements of different joint positions in the upper extremity with experienced certified hand therapist (CHT)
Weaknesses - Goniometer measurements were not taken in a position besides seated - The size of smartphones was not a controlled factor, which could affect body strain and positioning - Study sample was mostly females and included only OT graduate students, which makes it harder to generalize findings. - Goniometer measurement can be subjective - Technique used to measure scapular protraction has not been empirically tested |
|
Silva, et al. [26]
|
Level 2B
Cross Sectional Cohort Study
Risk of Bias Low |
Participants N = 961 Adolescents (14-19 years old)
Inclusion Criteria - Must complete questionnaire - Must agree to assessment measures - Must not have musculoskeletal pain or injuries due to pregnancy, infectious, genetic, or traumatic disorders.
Study Setting - Brazil - Public Schools - Self-report questionnaire |
Intervention - Self-report questionnaire asking about electronic use and symptoms over past 6 months |
Electronic Use - Computers and electronic games - Age at start of use - Frequency of use -Weekly frequency of use - Weekly time of use -Time of use
Musculoskeletal Pain |
Significant Findings - 65.1% of adolescents reported some musculoskeletal pain, most commonly in thoracolumbar spine (46.9%) and upper limbs (20%). - Use of electronic devices averaged 583 minutes per week and was associated with pain in the cervical region and low back. - Symptoms were reported to interfere with performing activities of daily living including study tasks and sports. - Female gender was also found to be associated with pain in all assessed body regions. |
Strengths - Considers the unique impact of technological advancement in public schools and the effect of increased electronic use with musculoskeletal pain in adolescents.
Weaknesses - Poor generalizability to other populations. - Self-report bias possible. |
|
Szucs, et al. [27]
|
Level 2B
Cohort Study
Risk of Bias Low |
Participants N = 21 college students (18 females, 3 males) (21.1 ± 1.5-years-old)
Inclusion Criteria - Ability to achieve full cervical ROM - A BMI less than or equal to 25
Study Setting - Lab setting |
Intervention - Participants completed a five-minute questionnaire to better understand their daily technology usage. - Pictures of participants were taken in a neutral posture and as they performed standard tasks with 3 devices (mobile phone, tablet, laptop). - A mobile application calculated sagittal and coronal plane posture variables, which were compared between device conditions. |
Postural Assessment - Angular and translational postural variables in the coronal and sagittal planes were calculated using digitized landmarks on pictures taken with a device camera
Devices Used: - Desktop computer - Laptop computer - Tablet - Cellphone |
Significant Findings - Using a tablet, whether supported on a table or held in the hands, produced the greatest amount of forward head shift and neck flexion - Tablet use produced postures that were statistically different than many other devices and technology - regular use of this device may produce greater deleterious effects than regular use of other handheld devices/technology |
Strengths - Reliability and validity of The PostureScreen Mobile® Application with good outcomes
Weaknesses - Activities were being done in a lab with standardized methods - Collected data not long after they began completing the task - Investigators provided the laptop and tablet - Did not compare postures during technology use with posture during reading a typical book or completing a pencil and paper writing task - Participants were all right-hand dominant and healthy college students |
|
Tapanya, et al. [28]
|
Level 2B
Non Randomized (Quasi-Experimental)
Risk of Bias Low |
Participants N = 32 young adults (16 males, 16 females) (18-25 years of age)
Inclusion Criteria - Participants had at least 6 months of smartphone use and used their smartphone daily for at least two hours per day
Study Setting - Physical Therapy Laboratory - Faculty of Associated Medical Sciences - KhonKaen University, Thailand. |
Intervention - Participants were randomly put into 4 different neck postures, including 0, 15, 30, and 45 neck flexion angles with elector - Participants were instructed to stand with feet shoulder-width apart - Participants were instructed to text for 3 minutes
|
Pain Assessment - Participants were asked to rate their neck discomfort score on a visual analogue scale before and after all assessments were completed in each condition.
Relevant measures taken: - Electromyography of 4 neck muscle groups (left cervical erector spinae, right cervical erector spinae, left upper trapezius, right upper trapezius) -Neck kinematics directly observed via 2-dimensional video recording
Devices Used: - Smartphone |
Significant Findings - Using a smartphone in a greater flexed neck posture results in a larger biomechanical burden on kinematics, gravitational moment and neck muscle loading which may increase the risk of neck musculoskeletal discomfort and injuries. - An appropriate neck posture for operating a smartphone while standing is represented by 0 neck flexion. - To use a smartphone with a 0 neck flexion, the appropriate phone tilt and gaze angles were considered to be 76 ± 5 and 14 ± 5, respectively |
Strengths - A sample size power level calculated to be 90%
Weaknesses - Only a static posture was measured. - Unable to identify the most appropriate shoulder angle for smartphone use. - Smartphone texting tasks were relatively short. |
|
Level 2B
Longitudinal Study
Risk of Bias Low |
Participants N = 1691 adolescents (829M/862F) Age- 10-19 years
Inclusion Criteria - Use of MTSD - Enrolled in primary or secondary schools - parent/guardian consent
Study Setting - Singapore - Web-based questionnaire at school |
Intervention - Response to survey at baseline and at one year follow-up. |
Relevant Measures Taken - Prevalence of MTSD use at baseline - Length of MTSD use - Activities using MTSD - Location of musculoskeletal pain (Neck/shoulder, upper back, lower back, arms. wrist/hand)
Devices Tested -Mobile Touch Screen Devices (smartphone, tablet ) |
Significant findings - Baseline to one year follow-up 74% of participants reported neck/shoulder symptoms - Baseline prevalence of phone use associated with neck/shoulder and low back symptoms - Baseline tablet use associated with neck/shoulder, low back, and arm symptoms at follow-up - Bout length >1 hr of smartphone at baseline associated with neck/shoulder and upper back symptoms at follow up, while >1hr of tablet use at baseline associated with low back symptoms. - Participation in certain activities (social, games, watching videos, general use) on MTSD at baseline associated with musculoskeletal symptoms at follow-up - Multitasking on smartphones at baseline associated with neck/shoulder and arm symptoms have follow-up. - No relationship between MTSD use duration (hrs/day) per activity and musculoskeletal systems. - Changes in MTSD use between Baseline and follow up not associated with severity of symptoms. |
Strengths - First longitudinal study on associations between MTSD use and musculoskeletal symptoms and visual health outcomes among adolescents. - Large sample with high follow up. - Detailed measures across a whole week including obtaining use during weekdays and weekend days. - Adjustments for known confounding factors.
Weaknesses - Self-report measures may introduce recall bias and inaccuracy. - Lack of multiple time point measures of symptoms. - Prior use or lifetime exposure to MTSD was not measured, and cumulative exposure may increase risk for symptoms. - Some adolescents dropped out and those who were lost to the follow up have significantly higher smartphone use duration and prevalence of musculoskeletal symptoms; their omission likely made findings more conservative.
|
|
|
Young, et al. [37] |
Level 3B
Cohort Study
Risk of Bias Low |
Participants N = 15 Adults - (7M/8F) - (Mean age = 29)
Inclusion Criteria - Experience with tablet/ computer - No previous or current musculoskeletal conditions of head, neck, back, or UE
Study Setting - Simulated in lab |
Intervention 2 different tablets (either ipad or Xoom) with configurations in: - support of tablet - location (lap vs table) - software task (web browsing, email, games) - Hand use (dominant vs. non-dominant)
Control - This study had no control group, with all participants experiencing the intervention to see significant differences in use. |
Posture and Muscle Activity - Posture of shoulder (3-D kinematics) - Posture of trunk (infrared 3-D motion analysis system) - Wrist (bi-axial goniometer) - Muscle activity (EMG)
Devices Used: - Xoom Tablet - ipad Tablet |
Significant Findings - Mean wrist posture, wrist angle accelerations, and muscle activity varied significantly across configurations and dominant vs. nondominant hand - Significantly more ulnar deviation noted in wrist during software task of emailing email
Non-significant Findings - No significant difference between the 2 different tablets and the participant’s posture - Wrist radial deviation when holding tablet - Higher values of wrist extension noted when using tablet |
Strengths - Extensive statistical analysis
Weaknesses - Simulated nature of environment and tasks may change natural behavior - Measurements taken over one time frame for each intervention and did not note extended use. |
|
Key: M: Male; F: Female; OT: Occupational Therapy; OTD: Doctorate of Occupational Therapy; UE: Upper Extremity; MTSD: Mobile Touch Screen Devices; DM: Diabetes Mellitus; GMA: Gravitational Moment Arm; CCL: Centroid Cervical Lordosis. Citation: Table format adapted from American Occupational Therapy Association [17]. |
||||||