Modeling with renormalization group and randomization 2

... Vol 24 , no 5, pp 723 -730, 15 Apr 20 09 [20 ] A Tay, SC Yen, JZ Li et al., “Real-time Gait Monitoring for Parkinson Disease,” Int Conf on Control and Automation., pp 1796-1801, June 12- 14, 20 13 [21 ]... between Kinesia system assessments and clinical tremor scores in patients with essential tremor”, Movement Disorders, Vol 25 , no 12, pp 1938-1943, 15 Sep 20 10 109 [22 ] LaStayo PC, Wheeler DL, ”Reliability... of Singapore June 20 12 to May 20 13 InnovFest 20 14 by NUS Enterprise 14 to 16 April 20 14 106 Bibliography [1] Neil R Sims, Hakan Muyderman, “Mitochondria, oxidative metabolism and cell death in

4.3 Preliminary Testing

Preliminary testing of the application was done on a healthy elderly test subjectwho stays with her caretaker The aim of this test was to check the reliability of thesystem when it is left at a patient’s home and to investigate if the caretaker can helpthe patient to use the tele-rehabilitation system During the initial set-up of thesystem, the caretaker was trained to operate the system and a therapist prescribedexercises for the subject The system was left at the patient’s home for a week tocollect data It was reported that the system failed to capture data for some ofthe days Improvements are being made to the system in response to the feedbackgathered from this week of testing in order to make it more reliable and robust Theprocess of detecting if the system is set up properly is being automated, so that thecaretaker can be reminded of set up procedures if needed Step-by-step instructionsare also being added to guide the caretaker in the placement of the sensors as thisdi↵ers across exercises The system will be tested on another elderly patient using

a similar protocol after these improvements

A RCT is carried out with the primary goal of evaluating if using the rehabilitation system in the first three months after the initial discharge results

tele-in a greater functional recovery among stroke patients, as compared to usual care.This is a collaboration between NUS, AMK-THKH and SGH (SGH wards in BrightVision Hospital (BVH)) Within NUS, the medical aspects of the study is managed

by Saw Swee Hock School of Public Health (SSH) and the engineering aspects of thestudy is managed by Department of Electrical and Computer Engineering (ECE).Stroke patients from the participating hospitals are screened against the inclu-sion and exclusion criteria set out by sta↵ from SSH If these patients provide aninformed consent to participate in the study, they are randomly allocated to one

79

of 2 possible groups: Control and Intervention The patients in the control groupwill be advised to attend supervised physiotherapy sessions at a day rehabilitationcentre nearby The patients in the intervention group will be provided with thetele-rehabilitation system at home for the first three months, with the therapistcontacting them once a week via FaceTime The FaceTime sessions will be used toreview the patients’ progress for the week and to adjust the difficulty of the exercises

if necessary Both control and intervention group patients undergo an assessment

at before the start of the study, at the 3rd month and the 6th month The assessorsperforming these assessments will be blinded to whether the patient is in control orintervention group The results of the assessments will be used to test the hypoth-esis of the study Patients in either group can choose to discontinue participation

in the study at anytime The focus of the engineering work of this project is onthe intervention group as the control group does not use the tele-rehabilitation sys-tem This section of the chapter will review the data collected from the first threepatients and therapists and the usability issues associated with the system

Patient Name A002 B002 B003 A005

Week 1 SMTWTFS SMTWTFS SMTWTFS SMTWTFSWeek 2 SMTWTFS SMTWTFS SMTWTFS SMTWTFSWeek 3 SMTWTFS SMTWTFS SMTWTFS SMTWTFS

Table 4.1: This table provides an overview of days of activity within the first 3weeks of intervention Days that are in bold are those days during which the first

4 patients performed exercises

Table 4.4 shows how frequently the first few patients are using the system inthe first few weeks of intervention This is a indication of whether the caretakersand the patients are comfortable with operating the system by themselves Thisalso indicates if the training with the tele-rehabilitation system that the patientsreceived before their discharge was enough to equip them with confidence and skills

to operate the system on their own It is worth looking at multiple patients asnon-compliance to the exercise schedules may not necessarily signal unfamiliaritywith the system The patients are supposed to perform their prescribed exercises

80

everyday No falls or adverse events were reported for these patients during theweeks shown in Table 4.4 Table 4.4 shows that patient A002’s adherence to theexercise schedule is low, at 2 to 3 days a week It has been confirmed that A002’scaretaker is able to operate the system and the low adherence was probably due

to the other activities that patient A002 is involved in, such as working at a tory, watering plants and feeding birds A change of caretaker in the third week isalso a likely cause of low adherence There is also another issue that this patient

fac-is not very motivated on performing the exercfac-ises After comparing multiple tients’ levels of motivation, it has been determined that this maybe due to the waythe pre-discharge hospital training was conducted for A002 For the case of A002,the responsibility of ensuring that the patient regularly performs the exercises washanded to the caretaker The caretaker was taught how to operate the system withthe sensors For the subsequent patients in the intervention group, this responsibil-ity was assigned to the patient as we taught the patient how to operate the systemand wear the sensors The caretaker’s role would be to help out if necessary This,together with the low motivation and change of caretaker could account for thepatient’s low adherence

pa-81

This chapter will describe the essential and possible future developments orresearch work that can be pursued on this tele-rehabilitation system One obviousfuture work is to finish up the RCT study Besides the study’s primary medicalaim, the study is also instrumental in getting feedback about the system from theusers This feedback allows for changes to be made and tested immediately, as long

as it does not a↵ect the fairness of the study The iteration process of making and

82

testing changes will enable developers to address usability issues Some of theseissues (mostly usability related) and other feedback points with the patient side ofthe system are listed below:

• Clarity of demonstration videos The demonstration videos for exercisesinvolving small limbs such as wrist and ankle do not show the small limbmovement clearly This causes the patient to lean forward and focus on thevideo before starting to perform the movements on his/her own

• Independent demonstration videos Both patients and therapist seem to

be keen on having access to demonstration videos On the patient side, thevideos should be accessible without having to turn the sensors on This feature

is requested with the aim of familiarising the patients with the exercises andfamiliarising the therapists with the demonstration videos provided to thepatients

• Detection of correct position This involves detecting if the patient isindeed in the starting position required for each of the exercises

• Detection of incorrect sensor orientation This involves detecting if thesensor is oriented wrongly on the limb by checking the orientation This can

be used to compensate for the incorrect orientation or to remind the patient

of the correct orientation for the limb sensor

• Video upload This is an issue faced by many of the patients and therapists.The patients tend to put their iPads to sleep before the videos of the variousexercises get uploaded This results in the therapist being unable to view thelater videos

The issues (mostly usability related) with the therapist side of the system are:

• Get therapist application to work on iPad 4 iPad 2 is being phased out

by apple and it would be increase the portability of the application if it were

83

to be able to work on retina iPads Currently this is to possible to due to theincompatibility of the compiled core-plot library.

• Alternative method for graph-plotting The graphs are taking a long time

to appear as time is required to communicate with the server, process the dataand draw the graphs This process can be sped up through pre-generatingthe graphs and storing them on the server The therapist application will justfetch and display the images in this case

• Independent demonstration videos Both patients and therapist seem to

be keen on having access to demonstration videos This feature is requestedwith the aim of familiarising the patients with the exercises and familiarisingthe therapists with the demonstration videos provided to the patients

• E-mailing of notes and activity prescriptions Therapists seem to find ituseful to have a copy of the call notes and the changes made to the activityprescriptions

• Compliance analytics A summary of how frequently a patient has formed the exercises, together with the number of repetitions performed may

per-be useful to the therapists

Another future work would be to expand the set of exercises o↵ered by the tem These include the integration of the ability to monitor functional movementssuch as tea-pouring, hair-combing and table-wiping

sys-Lastly, scalability of the system must be considered before allowing public use.The patient and therapist applications must be made available from the Apple’sApp Store, in order to aid mass distribution It would also help in distribution ifthe applications are ported over to Android The scalability of the server softwarecan be ensured by moving over to easily scalable cloud services, such as Amazonweb services, where the compute capacity and databases can adjust to growing

84

demands On the hardware side, scalability would involve getting a manufacturerand distributer to make and bring the sensor sets to the customers The systemmay also require regulatory approval for use in medical institutions.

85

Appendix A

Database Tables

Table A.1: Therapist information table (therapist info) to store data identifyingthe therapists

Column Name mySQL field name Variable type

Therapist ID therapist id mediumint unsigned not null

Username user name varchar(40)

Password password blob

Facetime e-mail facetime email varchar(100)

86

Table A.2: Patient information table (patient info) tostore data identifying the patients

Column Name MySQL Field Name Variable Type Remarks

Patient ID patient id mediumint unsigned not null Starts from 1, auto-incrementsSMF trial number trial no varchar(7)

Patient initials or name patient initials char(255)

Facetime e-mail facetime email Varchar(200)

Date of registration dor date

ID of therapist assigned to patient therapist id mediumint unsigned not null Foreign key from table 1.Paralysed side paralysed side varchar(1) L or R

iPad number iPad number smallint(6)

Site of recruitment (SGH/BVH) site text

Whether the initial log is done or not initial text

Whether the 3 month log is done or not three month text

Whether the 6 month log is done or not six month text

Continued on next page

Table A.2 – continued from previous pageColumn Name MySQL Field Name Variable Type Remarks

Date of birth dob date

Gender of patient gender text

Age of patient age text

The ward in which the patient was

admit-ted

ward text

Ethnic group that the patient belongs to ethnic text

Marital status of the patient married text

Date of stroke onset on the patient dos date

Date of admission of the patient to

hospi-tal

doa date

Date of discharge of patient from hospital,

after initial treatment

dod date

Date that the patient started the exercise

study at home AKA date of intervention

startdate date

Continued on next page

Table A.2 – continued from previous pageColumn Name MySQL Field Name Variable Type Remarks

Date that the patient ended the study

This is often 12 to 13 weeks after the

start-date

enddate date

The type of stroke that the patient has stroketype text

Whether the stroke is recurrent in the

pa-tients family or it is recurrent

strokehist text

Modified ranking scale scale text

Whether the patient has cardiovascular

disease

cardio text

Whether the patient has hypertension hyper text

Whether the patient has peripheral

vascu-lar disease

vascul text

Whether the patient has Chronic

Obstruc-tive Lung Disease

lung text

Continued on next page

Table A.2 – continued from previous pageColumn Name MySQL Field Name Variable Type RemarksWhether the patient has Arthritis or Mus-

Table A.3: Strengthening Activity Lists table to store information about allstrengthening activity

Column Name Variable type Remarks

activity id tinyint unsigned not null Starts from 1, auto-incrementsactivity name varchar(40)

upper or lower limb varchar(6)

joint ex id tinyint

joint id tinyint

min angle smallint

max angle smallint

Table A.4: Strengthening Progression levels table to store information about theprogression levels of strengthening exercises

Column Name Variable type Remarks

progression level ID tinyint unsigned not null Starts from 1, auto-incrementsprogression level name varchar(100)

theraband color varchar(10) Null if not applicable

Table A.5: Strengthening activity selection table to store the therapists’ ening activity prescriptions for the patients

strength-Column Name Variable type Remarks

activity ID tinyint unsigned not null Foreign Key from table 3patient ID mediumint unsigned not

null

Foreign Key from table 2

Is selected for patient boolean Yes or no, Default: notarget angle smallint unsigned Default: null

progress level ID tinyint unsigned not null Foreign key from table 4

Default: 1 Cannot be null.remarks text Default: null

91

Table A.6: Strengthening activity record table to store the patients’ strengtheningactivity repetitions.

Column Name Variable type Remarks

time of activity timestamp automatically set to

row update/ tion time

inser-patient ID mediumint unsigned

max time int unsigned default value of 0

minimum angle smallint default -500

maximum angle small int unsigned default -500

file name varchar(50) Null if file storing not

successful

Table A.7: Compliance form data table to store the number of minutes of exerciseper day for every patient

Column Name Variable type Remarks

trial number smallint unsigned not null foreign key from table 10day number tinyint unsigned

no of mins with therapist smallint unsigned

no of mins doing exercise smallint unsigned

92

Table A.8: HR-BP record table to store the heart-rate and blood pressure values

of patients

Column Name Variable type Remarks

time of entry timestamp

patient id mediumint unsigned not null foreign key form table 2

therapist id mediumint unsigned not null foreign key form table 1

bp systolic tinyint unsigned not null

bp diastolic tinyint unsigned not null

heart rate tinyint unsigned not null

Table A.9: FaceTime calls record table to record the date and time of calls made

by the therapist, and call notes

Column Name Variable type Remarks

button press time datetime not null

therapist id mediumint not null foreign key from table 1

patient id mediumint not null foreign key from table 2

end of call time datetime

last updated time timestamp

was app killed tinyint not null

notes by therapist text

Table A.10: iPad record table to keep track of the various iPads that are under use.Column Name Variable type

iPad number smallint not null auto increment

device serial number varchar(20) null

version number varchar(4) null

vendor id varchar(40) null

Table A.11: Seen graphs record table to keep track of the graph-pages that havebeen seen by each therapist

Column Name Variable type

patient id mediumint unsigned not null

therapist id mediumint unsigned not null

activity id tinyint unsigned not null

latest date of exercise in graph date null

number of sets on latest date smallint null

93

Appendix B

Exercises for testing of sensor accuracy

Table B.1: Outline of ROM assessment exercises