Counting number of events within an interval by group and identifying runs of events

[Mary Malakellis]

Hi,

I am working with an emergency department dataset with 12 years of data. I would like to identify patients that have 5 or more emergency department visits in 12-month rolling intervals. To do this I am using rangestat in Stata 17 as follows:

bysort patient_id: gen total_count = _n
rangestat (count) rolling_count = total_count, interval(arrival_date -364 0) by(patient_id)

I would like to flag patients with runs of 5 or more visits in rolling 12-month intervals. For example, for patient 6 I would like the run of 5 or more visits to be flagged as 1 and for patient 6 where the run is <5 to be 0 as seen in wanted.

I have a feeling I may need to use rangerun as follows:

rangerun total_count, interval(arrival_date -364 0) by(patient_id)

However, I am uncertain of how to define/construct a program with the commands.

There have been a number of threads if have reviewed as suggested such as https://www.statalist.org/forums/for...evious-5-years, but am still unable to follow (I am very new to coding).

Thank you, Mary

* Example generated by -dataex-. For more info, type help dataex
clear
input int patient_id float(arrival_date total_count) double rolling_count float wanted
 1 17958  1 1 0
 1 18223  2 2 0
 1 18527  3 2 0
 1 19075  4 1 0
 1 19757  5 1 0
 1 20737  6 1 0
 1 20743  7 2 0
 2 18031  1 1 0
 2 19416  2 1 0
 2 19676  3 2 0
 3 17955  1 1 0
 4 21813  1 1 0
 5 18089  1 1 0
 5 19535  2 1 0
 5 20826  3 1 0
 6 18368  1 1 1
 6 18385  2 2 1
 6 18406  3 3 1
 6 18430  4 4 1
 6 18494  5 5 1
 6 18552  6 6 1
 6 18558  7 7 1
 6 18563  8 8 1
 6 20429  9 1 0
 6 20891 10 1 0
 6 21605 11 1 0
 6 23006 12 1 0
 7 20869  1 1 0
 7 22486  2 1 0
 8 18631  1 1 0
 8 18632  2 2 0
 8 19548  3 1 0
 8 19549  4 2 0
 8 20689  5 1 0
 8 20990  6 2 0
 9 18413  1 1 0
 9 19453  2 1 0
 9 19469  3 2 0
 9 20059  4 1 0
 9 20195  5 2 0
 9 20354  6 3 0
 9 20739  7 1 0
 9 20754  8 2 0
 9 20825  9 3 0
 9 20843 10 4 0
10 21158  1 1 0
10 21184  2 2 0
10 21205  3 3 0
10 21656  4 1 1
10 21858  5 2 1
10 21910  6 3 1
10 21946  7 4 1
10 21976  8 5 1
end
format %td arrival_date

[Nick Cox]

Counting the total count sounds like double counting, but it isn't what rangestat (from SSC) is doing here, as is shown by counting an indicator 1 for each visit.

That said, I find your new problem harder to pin down. Each observation can be part of several overlapping spells that each have the same number of visits. And each patient can have two or more disjoint spells that have the same number of visits. Below is one approach, but far from the only possibility. It doesn't match your example.

gen visit = 1 
rangestat (count) nvisits = visit, interval(arrival_date -364 0) by(patient_id)
assert nvisits == rolling_count 

* new stuff 
bysort patient_id : egen max = max(nvisits)  
gen whenmax = nvisits == max 
rangestat (max) WANTED = whenmax, interval(arrival_date 0 364) by(patient_id)

[Mary Malakellis]

Thank you for the response.

Your syntax does not quite identify the runs of >=5 ED visits as desired. I will try to better articulate my problem. I am trying to determine the number of ED presentations that represent a frequent presentation, that is, >=5 visits/12 months

1. Attempt 1: Determine number of frequent presentations in calendar year. Problem: Patients may have >=5 presentations across the boundary of 2 years.

2. Attempt 2: Determine number of frequent presentations in 12 month blocks identifying the index case for each patient and using that as the starting point for the first 12 month block. Problem: As above, patients may have >=5 presentations across the boundary of 2 blocks. However, this is 'more accurate' than attempt 1.

3. Attempt 3 (current): Determine the number of frequent presentations in rolling 12 month blocks and identify runs of >=5 presentations. As you have correctly identified this means there are overlapping spells and there are disjointed spells.

Ultimately, I would like to determine the number of ED presentations that represent frequent presentations per calendar year irrespective if the run of >=5 presentations occurs over the boundary of 2 years. The frequent presentations will be counted only in the year for which they occur.

I send below a more complicated patient (#44) that has overlapping spells. For this patient - my attempt 2 solution does not capture all the runs of >=5 presentations.

Attempt 2 Syntax:
gen visit = 1

gen block = arrival_date
bysort patient_id (arrival_date visit): replace block = block[_n-1] if block < block[_n-1] + 365 & _n>1
format %td block

bysort patient_id block: gen block2 = _n>1
bysort patient_id block: gen freq_count= _N

generate freq_present_5 = .
replace freq_present_5 = 1 if freq_count >= 5
replace freq_present_5 = 0 if freq_count <=4
label define freq_present_5 1 ">=5 presentations" 0 "<=4 presentations"
label values freq_present_5 freq_present_5

Attempt 3 Syntax:

rangestat (count) nvisits = visit, interval(arrival_date -364 0) by(patient_id)

sort patient_id arrival_date

If you compare freq_present_5 to nvisits you will see that my attemp2 syntax misses flagging 1 date as a frequent presentation as identified by nvisits.

Thank you.

* Example generated by -dataex-. For more info, type help dataex
clear
input int patient_id float(arrival_date visit block block2 freq_count freq_present_5) double nvisits
43 18181 1 18181 0  2 0  1
43 18465 1 18181 1  2 0  2
43 19172 1 19172 0  1 0  1
43 20707 1 20707 0  1 0  1
43 21437 1 21437 0  3 0  1
43 21443 1 21437 1  3 0  2
43 21456 1 21437 1  3 0  3
44 17918 1 17918 0  4 0  1
44 18002 1 17918 1  4 0  2
44 18078 1 17918 1  4 0  3
44 18271 1 17918 1  4 0  4
44 18355 1 18355 0  2 0  4
44 18574 1 18355 1  2 0  3
44 18875 1 18875 0  6 1  2
44 19144 1 18875 1  6 1  2
44 19150 1 18875 1  6 1  3
44 19167 1 18875 1  6 1  4
44 19168 1 18875 1  6 1  5
44 19197 1 18875 1  6 1  6
44 19309 1 19309 0  9 1  6
44 19334 1 19309 1  9 1  7
44 19393 1 19309 1  9 1  8
44 19460 1 19309 1  9 1  9
44 19510 1 19309 1  9 1  9
44 19528 1 19309 1  9 1  9
44 19547 1 19309 1  9 1  8
44 19555 1 19309 1  9 1  9
44 19607 1 19309 1  9 1  9
44 19685 1 19685 0  9 1  9
44 19773 1 19685 1  9 1  8
44 19774 1 19685 1  9 1  9
44 19791 1 19685 1  9 1 10
44 19792 1 19685 1  9 1 11
44 19889 1 19685 1  9 1 10
44 19961 1 19685 1  9 1  8
44 20018 1 19685 1  9 1  8
44 20038 1 19685 1  9 1  9
44 20121 1 20121 0 10 1  9
44 20124 1 20121 1 10 1 10
44 20206 1 20121 1 10 1  7
44 20215 1 20121 1 10 1  8
44 20217 1 20121 1 10 1  9
44 20256 1 20121 1 10 1  9
44 20262 1 20121 1 10 1 10
44 20328 1 20121 1 10 1 10
44 20414 1 20121 1 10 1  9
44 20415 1 20121 1 10 1 10
44 20502 1 20502 0  9 1  9
44 20569 1 20502 1  9 1 10
44 20575 1 20502 1  9 1 10
44 20591 1 20502 1  9 1  9
44 20592 1 20502 1  9 1 10
44 20605 1 20502 1  9 1 11
44 20617 1 20502 1  9 1 12
44 20624 1 20502 1  9 1 12
44 20782 1 20502 1  9 1  9
44 20888 1 20888 0  2 0  9
44 21038 1 20888 1  2 0  3
44 21292 1 21292 0  4 0  2
44 21326 1 21292 1  4 0  3
44 21391 1 21292 1  4 0  4
44 21513 1 21292 1  4 0  4
44 21850 1 21850 0  4 0  2
44 22082 1 21850 1  4 0  2
44 22117 1 21850 1  4 0  3
44 22176 1 21850 1  4 0  4
44 22485 1 22485 0  3 0  2
44 22486 1 22485 1  3 0  3
44 22560 1 22485 1  3 0  3
44 22891 1 22891 0  3 0  2
44 22918 1 22891 1  3 0  3
44 22925 1 22891 1  3 0  3
45 22414 1 22414 0  1 0  1
end
format %td arrival_date
format %td block
label values freq_present_5 freq_present_5
label def freq_present_5 0 "<=4 presentations", modify
label def freq_present_5 1 ">=5 presentations", modify

[Nick Cox]

Sorry, but this is now more complicated than I want to try to understand. The mix of rolling 12-month periods and calendar years seems awkward. The advice you don't want at this point is that it's not only a matter of coding this, it has to be explained to your readers. Conversely, if this is what someone is asking for, that is a different story.

Someone else may be able to help more.

Pedantically, "disjoint" refers to sets that don't overlap. "disjointed" doesn't have the same meaning.

[Mary Malakellis]

Thank you for the prompt response. I appreciate the time taken to respond to my query and the many queries posted. I will have a rethink about my strategy.

I have noted the incorrect use of disjointed.

[Mary Malakellis]

I wish to write one clarifying point for any readers following this thread. The intention is to count the number of ED visits considered a frequent presentation as a proportion of the total ED visits for each calendar year. It is not intended to count the number of patients that present frequently. For example, if 1 patient presents 20 times in 1 calendar year, then each of those 20 visits are counted in the tally for that year (i.e., It is not counted as 1). This means for patients that visit the ED >=5 times over a boundary of two years the visits only get counted in the year in which they appear (again, not counting the patient as 1 in each of those years). I hope this now makes better sense.

[Mary Malakellis]

Hi,

I have come back to this post to try again as I am really stuck on this problem and have been for a long time.

I hope to add clarity to what I am seeking. I have emergency department data for about 12 years and I wish to identify patients that visit 5 or more times in 12 month rolling intervals. There are time gaps as some patients visit 1 time in a given 12 month interval, many times in a 12 month interval, or not at all for some 12 month intervals. I calculate the 12 month rolling intervals as follows:

rangestat (count) nvisits = edvisits, interval(arrival_date -364 0) by(unique_id)

I now wish to tag each run of 5 or more visits per unique_id as presented in wanted (manual entry). That is, 1 if the visit is part of a run of 5 or more visits or a 0 if the run is less than 5 visits.

Please note, I have developed my thinking since the previous posts and do not wish to use calendar years to count visits.

Thank you,

Mary

* Example generated by -dataex-. For more info, type help dataex
clear
input float case long unique_id int arrival_date float(year edvisits) double nvisits float want
 1 1 20074 2014 1  1 1
 2 1 20159 2015 1  2 1
 3 1 20186 2015 1  3 1
 4 1 20233 2015 1  4 1
 5 1 20403 2015 1  5 1
 6 1 20640 2016 1  2 1
 7 1 20807 2016 1  2 1
 8 1 20820 2017 1  3 1
 9 1 20828 2017 1  4 1
10 1 20831 2017 1  5 1
11 1 20853 2017 1  6 1
12 1 20923 2017 1  7 1
13 1 21229 2018 1  2 0
14 1 21559 2019 1  2 0
15 1 21696 2019 1  2 0
16 2 20042 2014 1  1 0
17 2 21147 2017 1  1 1
18 2 21418 2018 1  2 1
19 2 21427 2018 1  3 1
20 2 21440 2018 1  4 1
21 2 21445 2018 1  5 1
22 2 21459 2018 1  6 1
23 2 21461 2018 1  7 1
24 2 21462 2018 1  8 1
25 2 21516 2018 1  8 1
26 2 21540 2018 1  9 1
27 2 21684 2019 1 10 1
28 2 21687 2019 1 11 1
29 2 21839 2019 1  5 1
30 2 21879 2019 1  6 1
31 2 21954 2020 1  5 1
32 2 21955 2020 1  6 1
33 2 22020 2020 1  7 1
34 2 22026 2020 1  8 1
35 2 22158 2020 1  7 1
36 2 22159 2020 1  8 1
37 2 22229 2020 1  8 1
38 2 22410 2021 1  4 1
40 2 22446 2021 1  5 1
42 2 22621 2021 1  3 0
44 3 20043 2014 1  1 0
45 4 20043 2014 1  1 0
46 5 20043 2014 1  1 0
47 6 20043 2014 1  1 0
48 7 20043 2014 1  1 0
49 8 20043 2014 1  1 0
50 8 20075 2014 1  2 0
51 8 20241 2015 1  3 0
52 8 20242 2015 1  4 0
53 8 20418 2015 1  4 0
end
format %tddd-Mon-YY arrival_date

[Nick Cox]

I am not sure that I understand this. I can get close to reproducing your variable, but not in observation 6.

* Example generated by -dataex-. For more info, type help dataex
clear
input float case long unique_id int arrival_date float(year edvisits) double nvisits float want
 1 1 20074 2014 1  1 1
 2 1 20159 2015 1  2 1
 3 1 20186 2015 1  3 1
 4 1 20233 2015 1  4 1
 5 1 20403 2015 1  5 1
 6 1 20640 2016 1  2 1
 7 1 20807 2016 1  2 1
 8 1 20820 2017 1  3 1
 9 1 20828 2017 1  4 1
10 1 20831 2017 1  5 1
11 1 20853 2017 1  6 1
12 1 20923 2017 1  7 1
13 1 21229 2018 1  2 0
14 1 21559 2019 1  2 0
15 1 21696 2019 1  2 0
16 2 20042 2014 1  1 0
17 2 21147 2017 1  1 1
18 2 21418 2018 1  2 1
19 2 21427 2018 1  3 1
20 2 21440 2018 1  4 1
21 2 21445 2018 1  5 1
22 2 21459 2018 1  6 1
23 2 21461 2018 1  7 1
24 2 21462 2018 1  8 1
25 2 21516 2018 1  8 1
26 2 21540 2018 1  9 1
27 2 21684 2019 1 10 1
28 2 21687 2019 1 11 1
29 2 21839 2019 1  5 1
30 2 21879 2019 1  6 1
31 2 21954 2020 1  5 1
32 2 21955 2020 1  6 1
33 2 22020 2020 1  7 1
34 2 22026 2020 1  8 1
35 2 22158 2020 1  7 1
36 2 22159 2020 1  8 1
37 2 22229 2020 1  8 1
38 2 22410 2021 1  4 1
40 2 22446 2021 1  5 1
42 2 22621 2021 1  3 0
44 3 20043 2014 1  1 0
45 4 20043 2014 1  1 0
46 5 20043 2014 1  1 0
47 6 20043 2014 1  1 0
48 7 20043 2014 1  1 0
49 8 20043 2014 1  1 0
50 8 20075 2014 1  2 0
51 8 20241 2015 1  3 0
52 8 20242 2015 1  4 0
53 8 20418 2015 1  4 0
end
format %tddd-Mon-YY arrival_date

bysort unique_id (arrival_date): gen spell = sum(_n == 1 | nvisits <= nvisits[_n-1]) 

egen max_visits = max(nvisits), by(unique_id spell) 

gen wanted = max >= 5 

list, sepby(unique_id spell)

     +--------------------------------------------------------------------------------------------+
     | case   unique~d   arrival~e   year   edvisits   nvisits   want   spell   max_vi~s   wanted |
     |--------------------------------------------------------------------------------------------|
  1. |    1          1   17-Dec-14   2014          1         1      1       1          5        1 |
  2. |    2          1   12-Mar-15   2015          1         2      1       1          5        1 |
  3. |    3          1    8-Apr-15   2015          1         3      1       1          5        1 |
  4. |    4          1   25-May-15   2015          1         4      1       1          5        1 |
  5. |    5          1   11-Nov-15   2015          1         5      1       1          5        1 |
     |--------------------------------------------------------------------------------------------|
  6. |    6          1    5-Jul-16   2016          1         2      1       2          2        0 |
     |--------------------------------------------------------------------------------------------|
  7. |    7          1   19-Dec-16   2016          1         2      1       3          7        1 |
  8. |    8          1    1-Jan-17   2017          1         3      1       3          7        1 |
  9. |    9          1    9-Jan-17   2017          1         4      1       3          7        1 |
 10. |   10          1   12-Jan-17   2017          1         5      1       3          7        1 |
 11. |   11          1    3-Feb-17   2017          1         6      1       3          7        1 |
 12. |   12          1   14-Apr-17   2017          1         7      1       3          7        1 |
     |--------------------------------------------------------------------------------------------|
 13. |   13          1   14-Feb-18   2018          1         2      0       4          2        0 |
     |--------------------------------------------------------------------------------------------|
 14. |   14          1   10-Jan-19   2019          1         2      0       5          2        0 |
     |--------------------------------------------------------------------------------------------|
 15. |   15          1   27-May-19   2019          1         2      0       6          2        0 |
     |--------------------------------------------------------------------------------------------|
 16. |   16          2   15-Nov-14   2014          1         1      0       1          1        0 |
     |--------------------------------------------------------------------------------------------|
 17. |   17          2   24-Nov-17   2017          1         1      1       2          8        1 |
 18. |   18          2   22-Aug-18   2018          1         2      1       2          8        1 |
 19. |   19          2   31-Aug-18   2018          1         3      1       2          8        1 |
 20. |   20          2   13-Sep-18   2018          1         4      1       2          8        1 |
 21. |   21          2   18-Sep-18   2018          1         5      1       2          8        1 |
 22. |   22          2    2-Oct-18   2018          1         6      1       2          8        1 |
 23. |   23          2    4-Oct-18   2018          1         7      1       2          8        1 |
 24. |   24          2    5-Oct-18   2018          1         8      1       2          8        1 |
     |--------------------------------------------------------------------------------------------|
 25. |   25          2   28-Nov-18   2018          1         8      1       3         11        1 |
 26. |   26          2   22-Dec-18   2018          1         9      1       3         11        1 |
 27. |   27          2   15-May-19   2019          1        10      1       3         11        1 |
 28. |   28          2   18-May-19   2019          1        11      1       3         11        1 |
     |--------------------------------------------------------------------------------------------|
 29. |   29          2   17-Oct-19   2019          1         5      1       4          6        1 |
 30. |   30          2   26-Nov-19   2019          1         6      1       4          6        1 |
     |--------------------------------------------------------------------------------------------|
 31. |   31          2    9-Feb-20   2020          1         5      1       5          8        1 |
 32. |   32          2   10-Feb-20   2020          1         6      1       5          8        1 |
 33. |   33          2   15-Apr-20   2020          1         7      1       5          8        1 |
 34. |   34          2   21-Apr-20   2020          1         8      1       5          8        1 |
     |--------------------------------------------------------------------------------------------|
 35. |   35          2   31-Aug-20   2020          1         7      1       6          8        1 |
 36. |   36          2    1-Sep-20   2020          1         8      1       6          8        1 |
     |--------------------------------------------------------------------------------------------|
 37. |   37          2   10-Nov-20   2020          1         8      1       7          8        1 |
     |--------------------------------------------------------------------------------------------|
 38. |   38          2   10-May-21   2021          1         4      1       8          5        1 |
 39. |   40          2   15-Jun-21   2021          1         5      1       8          5        1 |
     |--------------------------------------------------------------------------------------------|
 40. |   42          2    7-Dec-21   2021          1         3      0       9          3        0 |
     |--------------------------------------------------------------------------------------------|
 41. |   44          3   16-Nov-14   2014          1         1      0       1          1        0 |
     |--------------------------------------------------------------------------------------------|
 42. |   45          4   16-Nov-14   2014          1         1      0       1          1        0 |
     |--------------------------------------------------------------------------------------------|
 43. |   46          5   16-Nov-14   2014          1         1      0       1          1        0 |
     |--------------------------------------------------------------------------------------------|
 44. |   47          6   16-Nov-14   2014          1         1      0       1          1        0 |
     |--------------------------------------------------------------------------------------------|
 45. |   48          7   16-Nov-14   2014          1         1      0       1          1        0 |
     |--------------------------------------------------------------------------------------------|
 46. |   49          8   16-Nov-14   2014          1         1      0       1          4        0 |
 47. |   50          8   18-Dec-14   2014          1         2      0       1          4        0 |
 48. |   51          8    2-Jun-15   2015          1         3      0       1          4        0 |
 49. |   52          8    3-Jun-15   2015          1         4      0       1          4        0 |
     |--------------------------------------------------------------------------------------------|
 50. |   53          8   26-Nov-15   2015          1         4      0       2          4        0 |
     +--------------------------------------------------------------------------------------------+

[Mary Malakellis]

Hi,

Thank you for your help.

Obs 6 is correctly identified as the second spell using your syntax, however, obs 6 also forms part of the third spell which is not picked up in this syntax. In the data below I have included each spell as a separate variable (a visual representation of nvisits). This shows obs 6 belongs to spell2 (which does not meet the criteria of >=5) and to spell3 (which does meet the criteria of >=5). As, such it should be 1 not 0.

Mary

* Example generated by -dataex-. For more info, type help dataex
clear
input float case long unique_id int arrival_date float(year edvisits) double nvisits float(spell1 spell2 spell3 spell4 spell5 spell6 spell7 spell8 spell9 spell10 want spell max_visits wanted)
 1 1 20074 2014 1  1 1 .  . . . . . . . . 1 1  5 1
 2 1 20159 2015 1  2 2 .  . . . . . . . . 1 1  5 1
 3 1 20186 2015 1  3 3 .  . . . . . . . . 1 1  5 1
 4 1 20233 2015 1  4 4 .  . . . . . . . . 1 1  5 1
 5 1 20403 2015 1  5 5 1  . . . . . . . . 1 1  5 1
 6 1 20640 2016 1  2 . 2  1 . . . . . . . 1 2  2 0
 7 1 20807 2016 1  2 . .  2 . . . . . . . 1 3  7 1
 8 1 20820 2017 1  3 . .  3 . . . . . . . 1 3  7 1
 9 1 20828 2017 1  4 . .  4 . . . . . . . 1 3  7 1
10 1 20831 2017 1  5 . .  5 . . . . . . . 1 3  7 1
11 1 20853 2017 1  6 . .  6 . . . . . . . 1 3  7 1
12 1 20923 2017 1  7 . .  7 1 . . . . . . 1 3  7 1
13 1 21229 2018 1  2 . .  . 2 1 . . . . . 0 4  2 0
14 1 21559 2019 1  2 . .  . . 2 1 . . . . 0 5  2 0
15 1 21696 2019 1  2 . .  . . . 2 . . . . 0 6  2 0
16 2 20042 2014 1  1 1 .  . . . . . . . . 0 1  1 0
17 2 21147 2017 1  1 . 1  . . . . . . . . 1 2  8 1
18 2 21418 2018 1  2 . 2  1 . . . . . . . 1 2  8 1
19 2 21427 2018 1  3 . 3  2 . . . . . . . 1 2  8 1
20 2 21440 2018 1  4 . 4  3 . . . . . . . 1 2  8 1
21 2 21445 2018 1  5 . 5  4 . . . . . . . 1 2  8 1
22 2 21459 2018 1  6 . 6  5 . . . . . . . 1 2  8 1
23 2 21461 2018 1  7 . 7  6 . . . . . . . 1 2  8 1
24 2 21462 2018 1  8 . 8  7 . . . . . . . 1 2  8 1
25 2 21516 2018 1  8 . .  8 1 1 . . . . . 1 3 11 1
26 2 21540 2018 1  9 . .  9 2 2 . . . . . 1 3 11 1
27 2 21684 2019 1 10 . . 10 3 3 1 . . . . 1 3 11 1
28 2 21687 2019 1 11 . . 11 4 4 2 . . . . 1 3 11 1
29 2 21839 2019 1  5 . .  . 5 5 3 1 . . . 1 4  6 1
30 2 21879 2019 1  6 . .  . . 6 4 2 1 . . 1 4  6 1
31 2 21954 2020 1  5 . .  . . . 5 3 2 . . 1 5  8 1
32 2 21955 2020 1  6 . .  . . . 6 4 3 . . 1 5  8 1
33 2 22020 2020 1  7 . .  . . . 7 5 4 . . 1 5  8 1
34 2 22026 2020 1  8 . .  . . . 8 6 5 . . 1 5  8 1
35 2 22158 2020 1  7 . .  . . . . 7 6 1 . 1 6  8 1
36 2 22159 2020 1  8 . .  . . . . 8 7 2 . 1 6  8 1
37 2 22229 2020 1  8 . .  . . . . . 8 3 . 1 7  8 1
38 2 22410 2021 1  4 . .  . . . . . . 4 1 1 8  5 1
40 2 22446 2021 1  5 . .  . . . . . . 5 2 1 8  5 1
42 2 22621 2021 1  3 . .  . . . . . . . 3 0 9  3 0
44 3 20043 2014 1  1 1 .  . . . . . . . . 0 1  1 0
45 4 20043 2014 1  1 1 .  . . . . . . . . 0 1  1 0
46 5 20043 2014 1  1 1 .  . . . . . . . . 0 1  1 0
47 6 20043 2014 1  1 1 .  . . . . . . . . 0 1  1 0
48 7 20043 2014 1  1 1 .  . . . . . . . . 0 1  1 0
49 8 20043 2014 1  1 1 .  . . . . . . . . 0 1  4 0
50 8 20075 2014 1  2 2 1  . . . . . . . . 0 1  4 0
51 8 20241 2015 1  3 3 2  . . . . . . . . 0 1  4 0
52 8 20242 2015 1  4 4 3  . . . . . . . . 0 1  4 0
53 8 20418 2015 1  4 . 4  . . . . . . . . 0 2  4 0
end
format %tddd-Mon-YY arrival_date