In the literature on sensing and monitoring, missing data is often treated as a nuisance. Statisticians have investigated many ways of filling gaps in a data set, depending on whether the data is missing completely randomly, whether the patterns of missing data can be predicted from other variables in the data set, or whether the patterns of missing data are regular, but hard to predict from existing data. In this talk, I argue that missing data is evidence for how people engage with sensors and devices. I outline the processes that result in the absence of data, and discuss how this contextual understanding can be used to improve interpretation and analytics. This talk is based on joint work with the Help4Mood team (http://help4mood.info) and Henry Potts, UCL, and Katarzyna Stawarz, Bristol, during an Alan Turing Institute Summer Programme Small Group.

1. partially joint work with Henry Potts and Katarzyna Stawarz; partially with the Help4Mood Consortium
The Hidden Stories
Behind Missing Data
Maria Wolters
University of Edinburgh
maria.wolters@ed.ac.uk
@mariawolters
http://www.slideshare.net/mariawolters

2. Overview
❖ What is Missing Data?
❖ The Story of Appropriation: Help4Mood
❖ The Story of Monitoring Work: Activity Trackers
❖ The Story of Hidden Missing Data: Electronic Health
Records

3. Missing Data
❖ informally: observations that we would like to be there,
or that should be there, but that are not
❖ Statistical treatment differs according to whether
missing data are
❖ completely random
❖ predictable from existing data
❖ not predictable from existing data

4. Goal
❖ Tell the hidden stories behind missing data by
understanding and describing data generation
processes
❖ qualitatively for deeper understanding
❖ quantitatively to feed into data analysis and
visualisation
❖ Current stage: setting out problem space in order to pick
appropriate models

5. The Appropriation of Help4Mood

6. depressioncomix.tumblr.com
Depression is a
change
relative to an
individual
baseline

7. Help4Mood: Supporting People with Depression
• daily monitoring
• of activity using
actigraph
• of mood, thought
patterns & psycho-
motor symptoms using
talking head GUI
• weekly one-page reports
to clinicians
Maria K. Wolters, Juan Martínez-Miranda, Soraya Estevez, Helen F. Hastie, Colin Matheson (2013). Managing Data in Help4Mood AMSYS ICST DOI: 10.4108/trans.amsys.
01-06.2013.e2

8. One Help4Mood per Day
❖ Users interact with Help4Mood once a day
❖ Sessions can be short, medium, or long
❖ Planning algorithm ensures
❖ different kinds of data are collected regularly, e.g.
❖ mood: every day
❖ speech: 1-3 times per week
❖ sessions are varied

9. Patient Focused Development
1. Actigraphy
2. Mood Tracker +
Actigraphy
3. Mood Tracker,
Thought Patterns,
Speech + Actigraphy
4. Mood Tracker,
Thought Patterns,
Behavioural Activation,
Relaxation,
Speech + Actigraphy
Case Studies
Pilot
Randomised
Controlled
Trial
People with
history of depression
People with
current depression

10. Pilot RCT Participants
❖ Participants with Major Depressive Disorder (SCID
diagnosed)
❖ Use Help4Mood for 4 weeks
❖ Background measures include demographics and
attitudes to computers
❖ (Pre/Post measures to establish change)
❖ Qualitative interviews at intake and debriefing for those
randomized to Help4Mood

11. Participants and Usage Patterns
❖ 18 in Romania, 7 in Scotland, 2 in Spain (EU Project)
❖ 14 treatment as usual (age 42 years +/- 10), 13
Help4Mood (age 35 +/- 12)
❖ None formally tracked or measured their mood before,
but some used introspection
❖ Half used it regularly, but that was not daily; instead, it
was 2-3 times per week. Why? Appropriation

12. What is Going On With Me?
The monitoring part helped me understand some things [. . .]
sometimes I did not realize how I felt that day, how happy I was
or how active I was. The system helped me observe these
things and also control them.
(RO14, female, 20–29)

13. Help4Mood as Coping
[. . .] when I felt very upset, very low, I preferred longer sessions [. . .]
because the session is pretty much focused on analyzing the thinking
patterns of depressed people, and those recovering from depression. It
gave me space to reflect on my own thoughts and feelings and gave me
alternative ways that I could resonate with.
(RO17, female, 30–39)
The most important thing for me was those relaxation exercises and I
think that it would have been very useful for the patient to have direct
access to them.
(RO03, female, 50–59)

14. It Doesn’t (Quite) Work This Way
http://imgarcade.com/1/depressed-stick-figure/
Well-Designed
Monitoring
Self-Help
Internet-Based
Therapy
+
http://www.acog.org/About-ACOG/ACOG-Departments/Long-Acting-Reversible-Contraception
https://www.osneybuyside.com/forget-big-data-just-collect-smart-data/
Protected
Peer Support
http://www.clipartsfree.net/small/3977-game-piece-group-clipart.html

15. It’s a complex adaptive system
http://www.thebolditalic.com/articles/3609-the-stick-figure-guide-to-kicking-depression
Individualised monitoring
based on what person has &
does
Coping and getting better:
• Twitter, exercise, kindness
• Friends
• Medications
• GP
Productive (!) self insight
and reflection

16. The Story of Monitoring Work

17. The Chore of Monitoring
If at all possible, it would be good not to have the same questions
every day; or even if the questions are the same, the phrasing
should be different. At some point it gets boring—I think this
could be changed.
(RO15, female, 30–39)

18. Self-Reflection is Hard Work
“This wasn’t very pleasant.
Because you don’t go to therapy every day.
You wouldn’t go every day; you would go maybe once a week
or two or three times maybe, but not every day.
It’s a bit too much to use it every day.”
(P01, Case Studies)

19. TRACKER
what
swimming
weightlifting
team sports
self-report
effort
e.g., Rooksby et al., CHI 2014; Lazar et al., CHI 2015

20. TRACKER
when
worried well
motivated for change
not tracking
during lazy days
what
swimming
weightlifting
team sports
stigma
self-report
effort if there is a need
e.g., Rooksby et al., CHI 2014; Lazar et al., CHI 2015

21. TRACKER
when
who job (e.g. nurses)
allergies
wrist anatomy
forgetting
to wear to bringto charge
worried well
techy
motivated for change
not tracking
during lazy days
what
swimming
weightlifting
team sports
style / fashion
stigma
self-report
effort if there is a need
e.g., Rooksby et al., CHI 2014; Lazar et al., CHI 2015

22. http://whitemenwearinggoogleglass.tumblr.com/image/51710586195
http://www.cio.com/article/2452759/health/wearable-techs-dilemma-too-much-data-not-enough-insight.html
Apple Watch Hermès

23. TRACKER
when
who job (e.g. nurses)
allergies
wrist anatomy
forgetting
to wear to bringto charge
worried well
techy
motivated for change
not tracking
during lazy days
device
breaks
no longer holds charge
lost / stolen
what
swimming
weightlifting
team sports
no Internet
style / fashion
stigma
self-report
effort
not synching
properly
if there is a need
e.g., Rooksby et al., CHI 2014; Lazar et al., CHI 2015

24. http://thoughtstipsandtales.com/2014/11/06/fitbit-fun-ten-months-later/
http://thoughtstipsandtales.com/2015/03/05/fabulous-fitbit-accessory-to-keep-the-clasp-from-opening/

25. TRACKER
when
who
forgetting is a function of
- user characteristics
- illness
- external stress
fit with identity
device
what
detailed user modeldevice model
connectivity model
fit with activities
effort required
to track activity
fit with ulterior need
(why tracking?)

26. TRACKER
when
who
forgetting is a function of
- user characteristics
- illness
- external stress
fit with identity
device
what
detailed user modeldevice model
connectivity model
fit with activities
effort required
to track activity
fit with ulterior need
(why tracking?)
All except for ulterior need and identity amenable to formal
/ quantitative modelling from appropriate discipline.

27. When looking at studies:
People will do things for research,
to benefit others, to be part of something bigger,
and maybe also to earn some money,
that they wouldn’t do for themselves.

28. The Story of Hidden Missing Data

29. Hiding in Electronic Health Records
❖ Data go dark because of quality issues in data entry and
management (e.g., Chan et al., 2014, Medical Care Research and Review)
❖ People go to the doctor when worried about something,
which increases likelihood of detection of other
problems.
❖ So does diabetes really increase your cancer risk, or is
your cancer more likely to be spotted in regular check
ups? (e.g, Badrick and Renehan, 2014, Eur J Cancer)

30. We can only detect what is being observed.
Does that mean regular screening and monitoring?
What is the human cost of a false positive?
What is the human cost of a false negative?
What are the numbers needed to treat?

31. Questions?
maria.wolters@ed.ac.uk
http://www.thebolditalic.com/articles/3609-the-stick-figure-guide-to-kicking-depression
Individualised monitoring
based on what person has &
does
Coping and getting better:
• Twitter, exercise, kindness
• Friends
• Medications
• GP
Productive (!) self insight
and reflection