The following examples are provided to show how to apply the Phoenix
Sepsis criteria within an SQLite data base. We will use the same example
data set sepsis from the R package for the following
examples. The examples are made possible by using the following R
packages
To help distinguish between R and sql code chunks, chunks outlined in blue are SQL and chunks with no outline are R chunks.
We start in R by setting up a local example SQL database with one
table sepsis
con <- dbConnect(drv = RSQLite::SQLite(), dbname = ":memory:")
dbWriteTable(conn = con, name = "sepsis", value = sepsis)We can easily query the table:
| pid | age | fio2 | pao2 | spo2 | vent | gcs_total | pupil | platelets | inr | d_dimer | fibrinogen | dbp | sbp | lactate | dobutamine | dopamine | epinephrine | milrinone | norepinephrine | vasopressin | glucose | anc | alc | creatinine | bilirubin | alt |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 0.06 | 0.75 | NA | 99 | 1 | NA | 199 | 1.46 | NA | 180 | 40 | 53 | NA | 1 | 1 | 1 | 1 | 0 | 0 | NA | NA | NA | 1.03 | NA | 36 | |
| 2 | 201.70 | 0.75 | 75.3 | 95 | 1 | 5 | both-reactive | 243 | 1.18 | 2.45 | 311 | 60 | 90 | 3.32 | 0 | 1 | 0 | 0 | 1 | 0 | 110 | 14.220 | 2.220 | 0.51 | 0.2 | 32 |
| 3 | 20.80 | 1.00 | 49.5 | NA | 1 | 15 | both-reactive | 49 | 1.60 | NA | 309 | 87 | 233 | 1.00 | 0 | 1 | 0 | 0 | 0 | 0 | 93 | 2.210 | 0.190 | 0.33 | 0.8 | 182 |
| 4 | 192.50 | NA | NA | NA | 0 | 14 | NA | 1.30 | 2.82 | 220 | 57 | 104 | NA | 0 | 0 | 0 | 0 | 0 | 0 | 110 | 3.184 | 0.645 | 0.31 | 8.5 | 21 | |
| 5 | 214.40 | NA | 38.7 | 95 | 0 | NA | 393 | NA | NA | NA | 57 | 101 | NA | 0 | 0 | 0 | 0 | 0 | 0 | NA | NA | NA | 0.52 | NA | NA | |
| 6 | 101.20 | 0.60 | 69.9 | 88 | 1 | 3 | both-reactive | 86 | 1.23 | 4.72 | 270 | 79 | 119 | 1.15 | 0 | 1 | 0 | 0 | 0 | 0 | 147 | 20.200 | 0.240 | 0.77 | 1.2 | 15 |
| 7 | 150.70 | 0.50 | NA | 31 | 1 | NA | 65 | 3.10 | NA | 94 | 11 | 14 | NA | 0 | 0 | 1 | 1 | 0 | 1 | NA | NA | NA | 1.47 | 1.7 | 3664 | |
| 8 | 159.70 | 0.30 | NA | 97 | 1 | 15 | both-reactive | 215 | 0.97 | 5.15 | 489 | 66 | 112 | NA | 0 | 0 | 0 | 0 | 0 | 0 | 100 | 3.760 | 1.550 | 0.58 | 0.5 | 50 |
| 9 | 176.10 | 0.65 | 51.0 | 82 | 1 | 3 | both-reactive | 101 | 1.08 | 7.71 | 456 | 51 | 117 | 8.10 | 0 | 0 | 1 | 1 | 1 | 1 | 264 | 8.770 | 3.600 | 1.23 | 21.1 | 151 |
| 10 | 6.60 | 0.80 | NA | 76 | 1 | 3 | both-reactive | 292 | NA | NA | NA | 58 | 84 | NA | 0 | 0 | 0 | 0 | 0 | 0 | 93 | 9.084 | 4.617 | 0.18 | NA | NA |
Respiratory
We need either create or code for the PF ratio, SF Ratio, invasive mechanical ventilation, and other respiratory support. In the construction of the PF ratio (pfr) and SF ratio (sfr) we replace missing values with value that will map to a score of zero. If the percent of inspiratory oxygen is greater than 0.21, that indicates the patient is on some form of respiratory support.
WITH respiratory AS (
SELECT
pid,
COALESCE(pao2 / fio2, 500) AS pfr,
COALESCE(IIF(spo2 <= 97, spo2 / fio2, 500), 500) AS sfr,
COALESCE(vent, 0) AS imv,
IIF(fio2 > 0.21 OR vent = 1, 1, 0) AS other_respiratory_support
FROM sepsis
)
SELECT *,
imv * (IIF(pfr < 100 OR sfr < 148, 1, 0) + IIF(pfr < 200 OR sfr < 220, 1, 0) ) +
other_respiratory_support * IIF(pfr < 400 OR sfr < 292, 1, 0) AS phoenix_respiratory_score
FROM respiratory;The above query was saved to an R data.frame called
resp and will be added to the virtual database for
inspection later.
dbWriteTable(conn = con, name = "respiratory", value = resp)Cardiovascular
Missing values correspond to zero points. The combination of age and mean arterial pressure will require both values to be know for possible points to be assessed.
WITH vars AS (
SELECT
pid,
COALESCE(dobutamine, 0) +
COALESCE(dopamine, 0) +
COALESCE(epinephrine, 0) +
COALESCE(milrinone, 0) +
COALESCE(norepinephrine, 0) +
COALESCE(vasopressin, 0) AS vasos,
lactate,
dbp + (sbp - dbp) / 3 AS map,
age
FROM sepsis
),
points AS (
SELECT *,
CASE WHEN vasos > 1 THEN 2
WHEN vasos > 0 THEN 1
ELSE 0 END AS vaso_points,
CASE WHEN lactate >= 11 THEN 2
WHEN lactate >= 5 THEN 1
ELSE 0 END AS lactate_points,
CASE WHEN (age >= 0 AND age < 1) AND (map < 17) THEN 2
WHEN (age >= 1 AND age < 12) AND (map < 25) THEN 2
WHEN (age >= 12 AND age < 24) AND (map < 31) THEN 2
WHEN (age >= 24 AND age < 60) AND (map < 32) THEN 2
WHEN (age >= 60 AND age < 144) AND (map < 36) THEN 2
WHEN (age >= 144 AND age <= 216) AND (map < 38) THEN 2
WHEN (age >= 0 AND age < 1) AND (map < 31) THEN 1
WHEN (age >= 1 AND age < 12) AND (map < 39) THEN 1
WHEN (age >= 12 AND age < 24) AND (map < 44) THEN 1
WHEN (age >= 24 AND age < 60) AND (map < 45) THEN 1
WHEN (age >= 60 AND age < 144) AND (map < 49) THEN 1
WHEN (age >= 144 AND age <= 216) AND (map < 52) THEN 1
ELSE 0 END AS map_points
FROM vars
)
SELECT *,
vaso_points + lactate_points + map_points AS phoenix_cardiovascular_score
FROM points;The results of the above query were saved behind the scene to an R
data.frame called card which will be added to
the SQLite database for inspection later.
dbWriteTable(conn = con, name = "cardiovascular", value = card)Coagulation
WITH points AS (
SELECT
pid,
CASE WHEN platelets < 100 THEN 1 ELSE 0 END AS plts,
CASE WHEN inr > 1.3 THEN 1 ELSE 0 END AS inr,
CASE WHEN d_dimer > 2 THEN 1 ELSE 0 END AS ddm,
CASE WHEN fibrinogen < 100 THEN 1 ELSE 0 END AS fib
FROM sepsis
)
SELECT *,
CASE WHEN plts + inr + ddm + fib >= 2 THEN 2
ELSE plts + inr + ddm + fib END AS phoenix_coagulation_score
FROM pointsThe above was stored as an R data.frame named
coag and will be added to the SQLite database for
inspection later.
dbWriteTable(conn = con, name = "coagulation", value = coag)Neurologic
WITH points AS (
SELECT
pid,
CASE WHEN gcs_total <= 10 THEN 1 ELSE 0 END AS gcs,
CASE WHEN pupil = "both-fixed" THEN 1 ELSE 0 END AS fixed_pupils
FROM sepsis
)
SELECT *,
CASE WHEN fixed_pupils = 1 THEN 2
WHEN gcs = 1 THEN 1
ELSE 0 END as phoenix_neurologic_score
FROM pointsThe above was stored as an R data.frame named
neuro and will be added to the SQLite database for
inspection later.
dbWriteTable(conn = con, name = "neurologic", value = neuro)Endocrine
SELECT
pid,
CASE WHEN glucose < 50 THEN 1
WHEN glucose > 150 THEN 1
ELSE 0 END AS phoenix_endocrine_score
FROM sepsisThe above was stored as an R data.frame named
endo and will be added to the SQLite database for
inspection later.
dbWriteTable(conn = con, name = "endocrine", value = endo)Immunologic
SELECT
pid,
CASE WHEN anc < 0.500 THEN 1 -- Recall expected units are 1000cells/mm³
WHEN alc < 1.000 THEN 1 -- Recall expected units are 1000cells/mm³
ELSE 0 END AS phoenix_immunologic_score
FROM sepsisThe above was stored as an R data.frame named
immu and will be added to the SQLite database for
inspection later.
dbWriteTable(conn = con, name = "immunologic", value = immu)Renal
SELECT
pid,
CASE WHEN (age >= 0 AND age < 1) AND (creatinine >= 0.8) THEN 1
WHEN (age >= 1 AND age < 12) AND (creatinine >= 0.3) THEN 1
WHEN (age >= 12 AND age < 24) AND (creatinine >= 0.4) THEN 1
WHEN (age >= 24 AND age < 60) AND (creatinine >= 0.6) THEN 1
WHEN (age >= 60 AND age < 144) AND (creatinine >= 0.7) THEN 1
WHEN (age >= 144 AND age <= 216) AND (creatinine >= 1.0) THEN 1
ELSE 0 END AS phoenix_renal_score
FROM sepsisThe above was stored as an R data.frame named
renal and will be added to the SQLite database for
inspection later.
dbWriteTable(conn = con, name = "renal", value = renal)Hepatic
SELECT
pid,
CASE WHEN bilirubin >= 4 THEN 1
WHEN alt > 102 THEN 1
ELSE 0 END AS phoenix_hepatic_score
FROM sepsisThe above was stored as an R data.frame named
hep and will be added to the SQLite database for inspection
later.
dbWriteTable(conn = con, name = "hepatic", value = hep)Phoenix
In R and python we provided wrappers for the total score. Here, we
will join together the results and get the totals. The results of the
following query are saved in an R data.frame called
phoenix_scores.
SELECT
respiratory.pid AS pid,
phoenix_respiratory_score,
phoenix_cardiovascular_score,
phoenix_coagulation_score,
phoenix_neurologic_score,
phoenix_respiratory_score + phoenix_cardiovascular_score +
phoenix_coagulation_score + phoenix_neurologic_score AS phoenix_sepsis_score,
IIF(phoenix_respiratory_score + phoenix_cardiovascular_score +
phoenix_coagulation_score + phoenix_neurologic_score >=2, 1, 0) AS phoenix_sepsis,
IIF(phoenix_respiratory_score + phoenix_cardiovascular_score +
phoenix_coagulation_score + phoenix_neurologic_score >=2 AND phoenix_cardiovascular_score > 0, 1, 0) AS phoenix_septic_shock
FROM respiratory
LEFT JOIN cardiovascular
ON respiratory.pid = cardiovascular.pid
LEFT JOIN coagulation
ON respiratory.pid = coagulation.pid
LEFT JOIN neurologic
ON respiratory.pid = neurologic.pid
phoenix_scores
## pid phoenix_respiratory_score phoenix_cardiovascular_score
## 1 1 0 2
## 2 2 3 2
## 3 3 3 1
## 4 4 0 0
## 5 5 0 0
## 6 6 3 1
## 7 7 3 4
## 8 8 0 0
## 9 9 3 3
## 10 10 3 0
## 11 11 3 3
## 12 12 1 0
## 13 13 0 0
## 14 14 2 2
## 15 15 3 3
## 16 16 0 2
## 17 17 2 2
## 18 18 3 2
## 19 19 2 2
## 20 20 0 1
## phoenix_coagulation_score phoenix_neurologic_score phoenix_sepsis_score
## 1 1 0 3
## 2 1 1 7
## 3 2 0 6
## 4 1 0 1
## 5 0 0 0
## 6 2 1 7
## 7 2 0 9
## 8 1 0 1
## 9 1 1 8
## 10 0 1 4
## 11 1 2 9
## 12 0 0 1
## 13 0 0 0
## 14 1 0 5
## 15 2 0 8
## 16 1 0 3
## 17 1 0 5
## 18 2 0 7
## 19 0 0 4
## 20 1 0 2
## phoenix_sepsis phoenix_septic_shock
## 1 1 1
## 2 1 1
## 3 1 1
## 4 0 0
## 5 0 0
## 6 1 1
## 7 1 1
## 8 0 0
## 9 1 1
## 10 1 0
## 11 1 1
## 12 0 0
## 13 0 0
## 14 1 1
## 15 1 1
## 16 1 1
## 17 1 1
## 18 1 1
## 19 1 1
## 20 1 1This output is the same as the output generated in R via:
phoenix_scores_from_r <-
phoenix(
# respiratory
pf_ratio = pao2 / fio2,
sf_ratio = ifelse(spo2 <= 97, spo2 / fio2, NA_real_),
imv = vent,
other_respiratory_support = as.integer(fio2 > 0.21),
# cardiovascular
vasoactives = dobutamine + dopamine + epinephrine + milrinone + norepinephrine + vasopressin,
lactate = lactate,
age = age,
map = dbp + (sbp - dbp)/3,
# coagulation
platelets = platelets,
inr = inr,
d_dimer = d_dimer,
fibrinogen = fibrinogen,
# neurologic
gcs = gcs_total,
fixed_pupils = as.integer(pupil == "both-fixed"),
data = sepsis
)
str(phoenix_scores)
## 'data.frame': 20 obs. of 8 variables:
## $ pid : int 1 2 3 4 5 6 7 8 9 10 ...
## $ phoenix_respiratory_score : int 0 3 3 0 0 3 3 0 3 3 ...
## $ phoenix_cardiovascular_score: int 2 2 1 0 0 1 4 0 3 0 ...
## $ phoenix_coagulation_score : int 1 1 2 1 0 2 2 1 1 0 ...
## $ phoenix_neurologic_score : int 0 1 0 0 0 1 0 0 1 1 ...
## $ phoenix_sepsis_score : int 3 7 6 1 0 7 9 1 8 4 ...
## $ phoenix_sepsis : int 1 1 1 0 0 1 1 0 1 1 ...
## $ phoenix_septic_shock : int 1 1 1 0 0 1 1 0 1 0 ...
str(phoenix_scores_from_r)
## 'data.frame': 20 obs. of 7 variables:
## $ phoenix_respiratory_score : int 0 3 3 0 0 3 3 0 3 3 ...
## $ phoenix_cardiovascular_score: int 2 2 1 0 0 1 4 0 3 0 ...
## $ phoenix_coagulation_score : int 1 1 2 1 0 2 2 1 1 0 ...
## $ phoenix_neurologic_score : int 0 1 0 0 0 1 0 0 1 1 ...
## $ phoenix_sepsis_score : int 3 7 6 1 0 7 9 1 8 4 ...
## $ phoenix_sepsis : int 1 1 1 0 0 1 1 0 1 1 ...
## $ phoenix_septic_shock : int 1 1 1 0 0 1 1 0 1 0 ...
identical(phoenix_scores[, -1], phoenix_scores_from_r)
## [1] TRUEPhoenix 8
For Phoenix 8 you need only four more organ dysfunction scores. The
result of the following query is saved as an R data.frame
called phoenix8_scores
SELECT
respiratory.pid AS pid,
phoenix_respiratory_score,
phoenix_cardiovascular_score,
phoenix_coagulation_score,
phoenix_neurologic_score,
phoenix_respiratory_score + phoenix_cardiovascular_score +
phoenix_coagulation_score + phoenix_neurologic_score AS phoenix_sepsis_score,
IIF(phoenix_respiratory_score + phoenix_cardiovascular_score +
phoenix_coagulation_score + phoenix_neurologic_score >=2, 1, 0) AS phoenix_sepsis,
IIF(phoenix_respiratory_score + phoenix_cardiovascular_score +
phoenix_coagulation_score + phoenix_neurologic_score >=2 AND phoenix_cardiovascular_score > 0, 1, 0) AS phoenix_septic_shock,
phoenix_endocrine_score,
phoenix_immunologic_score,
phoenix_renal_score,
phoenix_hepatic_score,
phoenix_respiratory_score + phoenix_cardiovascular_score +
phoenix_coagulation_score + phoenix_neurologic_score +
phoenix_endocrine_score + phoenix_immunologic_score +
phoenix_renal_score + phoenix_hepatic_score AS phoenix8_sepsis_score
FROM respiratory
LEFT JOIN cardiovascular
ON respiratory.pid = cardiovascular.pid
LEFT JOIN coagulation
ON respiratory.pid = coagulation.pid
LEFT JOIN neurologic
ON respiratory.pid = neurologic.pid
LEFT JOIN endocrine
ON respiratory.pid = endocrine.pid
LEFT JOIN immunologic
ON respiratory.pid = immunologic.pid
LEFT JOIN renal
ON respiratory.pid = renal.pid
LEFT JOIN hepatic
ON respiratory.pid = hepatic.pid
phoenix8_scores
## pid phoenix_respiratory_score phoenix_cardiovascular_score
## 1 1 0 2
## 2 2 3 2
## 3 3 3 1
## 4 4 0 0
## 5 5 0 0
## 6 6 3 1
## 7 7 3 4
## 8 8 0 0
## 9 9 3 3
## 10 10 3 0
## 11 11 3 3
## 12 12 1 0
## 13 13 0 0
## 14 14 2 2
## 15 15 3 3
## 16 16 0 2
## 17 17 2 2
## 18 18 3 2
## 19 19 2 2
## 20 20 0 1
## phoenix_coagulation_score phoenix_neurologic_score phoenix_sepsis_score
## 1 1 0 3
## 2 1 1 7
## 3 2 0 6
## 4 1 0 1
## 5 0 0 0
## 6 2 1 7
## 7 2 0 9
## 8 1 0 1
## 9 1 1 8
## 10 0 1 4
## 11 1 2 9
## 12 0 0 1
## 13 0 0 0
## 14 1 0 5
## 15 2 0 8
## 16 1 0 3
## 17 1 0 5
## 18 2 0 7
## 19 0 0 4
## 20 1 0 2
## phoenix_sepsis phoenix_septic_shock phoenix_endocrine_score
## 1 1 1 0
## 2 1 1 0
## 3 1 1 0
## 4 0 0 0
## 5 0 0 0
## 6 1 1 0
## 7 1 1 0
## 8 0 0 0
## 9 1 1 1
## 10 1 0 0
## 11 1 1 1
## 12 0 0 0
## 13 0 0 0
## 14 1 1 0
## 15 1 1 0
## 16 1 1 0
## 17 1 1 0
## 18 1 1 0
## 19 1 1 0
## 20 1 1 1
## phoenix_immunologic_score phoenix_renal_score phoenix_hepatic_score
## 1 0 1 0
## 2 0 0 0
## 3 1 0 1
## 4 1 0 1
## 5 0 0 0
## 6 1 1 0
## 7 0 1 1
## 8 0 0 0
## 9 0 1 1
## 10 0 0 0
## 11 0 1 0
## 12 0 0 0
## 13 0 0 0
## 14 0 0 0
## 15 0 1 1
## 16 0 0 0
## 17 0 0 0
## 18 0 1 1
## 19 0 1 0
## 20 0 0 0
## phoenix8_sepsis_score
## 1 4
## 2 7
## 3 8
## 4 3
## 5 0
## 6 9
## 7 11
## 8 1
## 9 11
## 10 4
## 11 11
## 12 1
## 13 0
## 14 5
## 15 10
## 16 3
## 17 5
## 18 9
## 19 5
## 20 3
phoenix8_scores_r <-
phoenix8(
# respiratory
pf_ratio = pao2 / fio2,
sf_ratio = ifelse(spo2 <= 97, spo2 / fio2, NA_real_),
imv = vent,
other_respiratory_support = as.integer(fio2 > 0.21),
# cardiovascular
vasoactives = dobutamine + dopamine + epinephrine + milrinone + norepinephrine + vasopressin,
lactate = lactate,
age = age, # Also used in the renal assessment.
map = dbp + (sbp - dbp)/3,
# coagulation
platelets = platelets,
inr = inr,
d_dimer = d_dimer,
fibrinogen = fibrinogen,
# neurologic
gcs = gcs_total,
fixed_pupils = as.integer(pupil == "both-fixed"),
# endocrine
glucose = glucose,
# immunologic
anc = anc,
alc = alc,
# renal
creatinine = creatinine,
# no need to specify age again
# hepatic
bilirubin = bilirubin,
alt = alt,
data = sepsis
)
str(phoenix8_scores)
## 'data.frame': 20 obs. of 13 variables:
## $ pid : int 1 2 3 4 5 6 7 8 9 10 ...
## $ phoenix_respiratory_score : int 0 3 3 0 0 3 3 0 3 3 ...
## $ phoenix_cardiovascular_score: int 2 2 1 0 0 1 4 0 3 0 ...
## $ phoenix_coagulation_score : int 1 1 2 1 0 2 2 1 1 0 ...
## $ phoenix_neurologic_score : int 0 1 0 0 0 1 0 0 1 1 ...
## $ phoenix_sepsis_score : int 3 7 6 1 0 7 9 1 8 4 ...
## $ phoenix_sepsis : int 1 1 1 0 0 1 1 0 1 1 ...
## $ phoenix_septic_shock : int 1 1 1 0 0 1 1 0 1 0 ...
## $ phoenix_endocrine_score : int 0 0 0 0 0 0 0 0 1 0 ...
## $ phoenix_immunologic_score : int 0 0 1 1 0 1 0 0 0 0 ...
## $ phoenix_renal_score : int 1 0 0 0 0 1 1 0 1 0 ...
## $ phoenix_hepatic_score : int 0 0 1 1 0 0 1 0 1 0 ...
## $ phoenix8_sepsis_score : int 4 7 8 3 0 9 11 1 11 4 ...
str(phoenix8_scores_r)
## 'data.frame': 20 obs. of 12 variables:
## $ phoenix_respiratory_score : int 0 3 3 0 0 3 3 0 3 3 ...
## $ phoenix_cardiovascular_score: int 2 2 1 0 0 1 4 0 3 0 ...
## $ phoenix_coagulation_score : int 1 1 2 1 0 2 2 1 1 0 ...
## $ phoenix_neurologic_score : int 0 1 0 0 0 1 0 0 1 1 ...
## $ phoenix_sepsis_score : int 3 7 6 1 0 7 9 1 8 4 ...
## $ phoenix_sepsis : int 1 1 1 0 0 1 1 0 1 1 ...
## $ phoenix_septic_shock : int 1 1 1 0 0 1 1 0 1 0 ...
## $ phoenix_endocrine_score : int 0 0 0 0 0 0 0 0 1 0 ...
## $ phoenix_immunologic_score : int 0 0 1 1 0 1 0 0 0 0 ...
## $ phoenix_renal_score : int 1 0 0 0 0 1 1 0 1 0 ...
## $ phoenix_hepatic_score : int 0 0 1 1 0 0 1 0 1 0 ...
## $ phoenix8_sepsis_score : int 4 7 8 3 0 9 11 1 11 4 ...
identical(phoenix8_scores[, -1], phoenix8_scores_r)
## [1] TRUE
dbDisconnect(con)