Refining the Process

Question:

• example might  be something like  ...  During _(event)_, does the body produce a chemical reaction that can be detected in the saliva or on the skin?

Research:

• this includes what has come before such as DAD information with dogs ability to detect BG levels...  what is already known that leads to to question/s

Hypothesis:

• during (an event) the body produces a chemical reaction that can be detected  1. on the skin 2. in the saliva by dogs trained to identify chemical changes in body.

Methods/Procedure:

• equipment  ( shakers, cans, wheel etc.)  
• protocol ... cleaning equipment etc...  housekeeping
• the test instrument ....  dog and dogs qualifications
• tests to be run 

Data:

• timelines, test records, flips  all evidence/documentation.

Observations:

• the product of the data and what it shows.

Conclusion:

• goes back to the hypothesis and addresses that in relation to the data and observations.  It then concludes the process and the results direct to next question and or further study.

Contamination and Cross Contamination

"I would contaminate EVERYTHING with the treat scent--the apparatus, random spots on the wall, my shoes, you name it. Rather than get into a futile effort to maintain sterile procedures, I'd let the dog learn that the only odor that pays is the target odor."

~Steve White

Observation Precedes Interpretation

"If I am busy interpreting behavior, I will often miss behavior that does not fit my interpretation. An observer should first of all observe. After duly recording what has gone on, then interpret, speculate, analyze, or do whatever, but, first, observe critically." 

~Bob Bailey 

SAP Testing Dogs Techniques

Video examples of each SAP testing dog working the wheel.

Dog 1

• indicates with nose
• filmed w/iPad on tripod
• uploaded via YouTube

Dog 2

• indicates with Down
• filmed with Flip sitting on floor
• uploaded via YouTube

Dog 3

• indicates w/nose
• filmed w/iPhone (horizontal) sitting on stool
• uploaded via YouTube

Dog 5

• indicates w/nose
• filmed w/iPhone (horizontal) on tripod
• uploaded via Vimeo

Dog 6

• indicates w/nose
• filmed w/Flip camera
• uploaded via Flipshare

The ICAN Study

under construction

http://media.wix.com/ugd/76ecfc_50bddb649464fa0f3e48e0970fb8680f.pdf

	ICAN	SAP
Purpose	To determine if ICAN trained Diabetes Alert Dogs (DADs) could alert to the scent of a low glycemic sample in the presence of normal and blank (control) samples in a controlled environment.
Description	In both studies, 4 separate wheels, containing samples from 4 separate patients were used and dog (Study 1: n=2, Study 2: n=4) replicated the search 2X/wheel. The placement of the samples was based on a pre-specified randomization scheme. Samples were placed in separate cups on a randomization device.  - One cup contained the low sample 3 contained euglycemic samples from the same patient 3 contained gauze w/o samples
Study Design	Each trained DAD (Study 1, n=2  Study 2, n=4) was tested on the same set of patient samples TWICE  (Study 1, k=4  Study 2, k=4) Each study was conducted with a different set of patients.
Method of Sample Capture and Preservation	Samples consisted of gauze squares wiped on the skin (back of the neck and forehead) and placed in a plastic zipper bag. Prior to sealing the zipper bag, the patient exhaled into the bag containing the gauze square.
Testing Participants	The Experiment Room had 5 individuals and 1 dog per trial. 4 individuals were blinded and 1 individual was unblinded to the study Blinded 1 dog handler 2 scorers 1 coordinator Unblinded Signaler Test Subject dog
Participant Roles	Dog Handler (blinded): -led the dog to and from the experiment room and gave the dog a treat for a correct response Scorers (blinded): -recorded the dog name, time, responses and sniffs by the dog Coordinator (blinded): watched the signaler for a correct response and told the dog handler to give the dog a treat upon a correct response Signaler (unblinded): -loaded the Scent Wheel with the patient samples and placed the Scent Wheel on the ground. During the study, the signaler notified the coordinator if the response was correct.
Scent Wheel Design	The Scent Wheel was designed to use 7 identical cups that hid the sample from sight, but allowed the scent to escape. The 7 cups included 1 low glycemic (low) 3 euglycemic (norm) 3 gauze only (blank samples) Each wheel included randomly placed samples from the same patient. The Scent Wheel was placed on the ground and rotated between trials (reps).
Scent Wheel Cleaning Procedures	The scent wheel and sample cups were cleaned before each study and between each trial Sample cups were washed thoroughly w/commercially available dish soap prior to the study commercially available cleaning wipes were used to ensure scent contamination was not present between trials
Testing Protocol	Each replicate had the same patient samples arranged in the same order, but the scent wheel was rotated to a  different orientation. Each replicate lasted a maximum of 30s or until a correct alert by the DAD.
Donor Acquisition Procedure
Terminology

A Scenting Test: Dog 1 and Dog 6

In an attempt to determine what it is that the dogs consider to be their Target Odor - the sample, residual odor or their own saliva on the shaker lids we decided to test that on the scent wheel.

Both Dog 1 and Dog 6 performed one Set of 10 reps.  The cap of the shaker containing the Target Odor was exchanged for the cap of a shaker containing a dummy sample and placed opposite each other on the wheel.

DOG ID #	1 and 6		TEST DATE		9/29/13
TEST	Experiment: Is dog detecting target odor or own saliva?
DONOR ID #			ORIGIN
SCENT TYPE (saliva/spit/sweat/other			SEIZURE TYPE
DATE of CAPTURE	Dog 1: 9/25/13 Dog 6: 9/24/13		TIME of CAPTURE (before/during/after)
SAMPLE STATE (cotton/gauze pad/clothing/other	Dog 1: dental cotton Dog 6:  Q tip		STORAGE CONTAINER (covered test tube/double bagged))		Both sample: double bagged in container
SAMPLE SUMMARY Dog 1	room temp LowBG Saliva Day5		SAMPLE SUMMARY Dog 2		room temp Sage Oil on Qtip Day 6
MEDICATIONS	Humalog (Dog 1)
SAMPLE SUMMARY	Both dogs had tested 2 completed sessions earlier in the day with their individual Target Odors
SAMPLE SUMMARY & DIFFERENCES	Dog 1 target in 1 cap in 4			Dog 6 target in cap in
INTENSITY (# of holes	5			11
CONTAINER TYPE	glass salt shaker			glass cheese shaker
PRESENTATION (hand/surface/wheel)	wheel			wheel
# of CONTAINERS	6	6	6	6	6	6
# of SCENT OPTIONS	1			1
REPS	10			10
% SUCCESS	70			80
COMMENTS	*1	*2	*3	*4	*5	*6
TARGET in #	1,1,1,1,1,  1,1,1,1,1  FP was always 4			X,X, FP was  always X
EASE OF FIND 1 Very easy     2 Easy   3 Normal 4 Difficult     5 Very difficult	3
TIME	2m 55s

Dog 1 had completed 2 Sessions plus 1 Ritual  (61 testing reps) during the day with this same Target Odor in the original shaker and cap.

Container #4 contained the neutral cotton and target cap was directly in front of the dog when the test began.  She indicated it.  I did not call her off, nor did I mark with the click, so she continued searching and as soon as she came to Container #1, holding the Target Odor, she indicated.

The second False Positive occurred when it was the first of the "two" containers that she came to, as was the final False Positive.  The False Positives were always #4.

The video makes this clear.  It is now almost as though there are two identical Target Odors on the wheel.

This same phenomenon occurred when Dog 2 tested human and canine seizure scent on the same wheel at the same time. 

 Dog 6 had completed 2 Sessions (60 testing reps) during the day with this same Target Odor in the original shaker and cap.

NOTE:  Waiting for narrative from Team 6.

Dog 5s trainer took the test even farther, giving it 6 different variations.

DOG ID #	5		TEST DATE		10/23/2013
TEST	Odor 3:  wheel
DONOR ID #			ORIGIN
SCENT TYPE (saliva/spit/sweat/other	Odor 3:		SEIZURE TYPE
DATE of CAPTURE	10/23/2013		TIME of CAPTURE (before/during/after)		am
SAMPLE STATE (cotton/gauze pad/clothing/other	1/2 Q-tip		STORAGE CONTAINER (covered test tube/double bagged))
SAMPLE CONDITION (frozen/refrigerated/room temp/other)	room temp		SAMPLE ID AGE when USED		Sample X Day X
MEDICATIONS	none
SAMPLE SUMMARY	tested @ roomt temp; stored XXXX
SAMPLE SUMMARY & DIFFERENCES
INTENSITY (# of holes	13
CONTAINER TYPE	glass salt shaker			glass salt shaker
PRESENTATION (hand/surface/wheel)	wheel
# of CONTAINERS	6	6
# of SCENT OPTIONS	1	1	1	1	1	1
REPS	10	10
% SUCCESS	70%	100%
COMMENTS	*1	*2	*3	*4	*5	*6
TARGET in #	3	5
EASE OF FIND 1 Very easy     2 Easy   3 Normal 4 Difficult     5 Very difficult	3	2
TIME	1:42	2:03

Dog 5:

*1.  Here scent was put in can 3.  Every lid had Boon's saliva on it.  70%

*2 Target scent in can 5.  Every lid has Boon's saliva on it.  100%

DOG ID #	5		TEST DATE		10/23/2013
TEST	Odor 3D cap test A
DONOR ID #			ORIGIN
SCENT TYPE (saliva/spit/sweat/other	Odor 3:		SEIZURE TYPE
DATE of CAPTURE	10/23/2013		TIME of CAPTURE (before/during/after)		9am
SAMPLE STATE (cotton/gauze pad/clothing/other	1/2 Q-tip		STORAGE CONTAINER (covered test tube/double bagged))
SAMPLE CONDITION (frozen/refrigerated/room temp/other)	room temp		SAMPLE ID AGE when USED		Sample X Day X
MEDICATIONS	none
SAMPLE SUMMARY	tested @ roomt temp; stored XXXX
SAMPLE SUMMARY & DIFFERENCES
INTENSITY (# of holes	13
CONTAINER TYPE	glass salt shaker			glass salt shaker
PRESENTATION (hand/surface/wheel)	wheel
# of CONTAINERS	6	6	6	6
# of SCENT OPTIONS	1	1	1	1	1	1
REPS	10	10	10
% SUCCESS	100%	90%	50%	90%
COMMENTS	*1	*2	*3	*4	*5	*6
TARGET in #	4	6	1	2
EASE OF FIND 1 Very easy     2 Easy   3 Normal 4 Difficult     5 Very difficult	2	2	4	2
TIME			2:51	3:18

VIDEOS/COMMENTS:

*1. Reminder set  Scent in one shaker, neutral in 5 shakers all clean shaker lids. No video 100%
*2. Set 2 moved scent to can 6.  No other changes. 90%
*3. Changed Saliva contaminated scent lid to shaker in can 3. Place a clean lid on scent shaker in can 1. Boon seemed to indicate whichever one came first.
Also, Trainer needs to click correct responses before notating!  Also helpful if a pocket is available for wrong choice treats to keep the count correct!

*4 Scent shaker with clean lid placed in can 2.  Contaminated lids now in cans 1 and 5.
Boon indicated can 5 once.  90% this trial

CONCLUSION:  We need to include switching caps into the teaching process to further strengthen the dogs understanding of her job - that she is only paid to identify the TARGET odor.