Table 1 Migraine-locked days.

From: A Markov chain method for counting and modelling migraine attacks

 

Raw headache diary data

Counts

Diary day:

29

30

31

32

33

34

35

36

37

\(\cdots \)

Days:

9

Migraine days:

4

Migraine:

 

M

 

M

  

M

M

 

\(\cdots \)

MLD:

1

Migraine attacks:

3

 

Data with imputed migraine locked days

Counts

Diary day:

29

30

31

32

33

34

35

36

37

\(\cdots \)

Days:

9

Migraine days:

5

Migraine:

 

M

M

M

  

M

M

 

\(\cdots \)

MLD:

0

Migraine attacks:

2

  1. Illustration of the Fill48-, or, imputation of migraine locked days- method for the accounting of migraine days and migraine attacks. The top panel shows an excerpt of a hypothetical migraine diary: the top row records the day’s number, the bottom row records migraine headache days. The day (numbered 31) is a migraine locked day: both the day preceding it and the day succeeding it is recorded with a migraine. The lower panel shows the diary post-imputation: the migraine-locked day is recorded as a migraine day (in italics). Both panels are accompanied with a table recording the relevant counts (from the excerpt from days 29–37): in the top diary—without imputed migraine locked days—we count a total of four migraine days distributed between 3 migraine attacks. In the bottom diary—with an imputed migraine locked day—we count a total of five migraine days, distributed between 2 migraine attacks. For the researcher focused on triggers, the raw data suggests importance of days 30, 32 and 35 as migraine onsets. If the Fill48 method is employed, only days 30 and 35 counts as migraine onset days.
Back to article page