Selection bias in clinical sciences: Difference between revisions
Selection bias in clinical sciences (view source)
Revision as of 10:44, 4 February 2024
, 4 months ago→{{header|Wren}}: Changed to Wren S/H
SqrtNegInf (talk | contribs) (Added Perl) |
m (→{{header|Wren}}: Changed to Wren S/H) |
||
| (2 intermediate revisions by 2 users not shown) | |||
Line 198:
degrees of freedom: 2
adjustment for ties: 0.417533
</pre>
=={{header|Nim}}==
{{trans|Python}}
{{trans|Wren}}
<syntaxhighlight lang="Nim">import std/[algorithm, math, random, sequtils, strformat, sugar]
type
TreatmentClass {.pure.} = enum Untreated, Irregular, Regular
Subject = object
cumDose: float
category: TreatmentClass
hadCovid: bool
const DoseForRegular = 100
func categorize(subject: var Subject) =
## Set treatment category based on cumulative treatment taken.
subject.category = if subject.cumDose == 0: Untreated
elif subject.cumDose >= DoseForRegular: Regular
else: Irregular
proc update(subject: var Subject; pCovid = 0.001;
pStart = 0.005; pDosing = 0.25; doses = @[3, 6, 9]) =
## Daily update on the subject to check for infection and randomly dose.
if not subject.hadCovid:
if rand(1.0) < pCovid:
subject.hadCovid = true
elif (subject.cumDose == 0 and rand(1.0) < pStart) or
(subject.cumDose > 0 and rand(1.0) < pDosing):
subject.cumDose += sample(doses).toFloat
subject.categorize()
func kruskalWallis(a, b, c: seq[bool]): float =
# Aggregate and sort.
let s = sorted(a & b & c)
# Find rank of first occurrence of "true".
let ix = s.find(true) + 1
# Calculate average ranks for "false" and "true".
let n = s.len
let arf = ix / 2
let art = (ix + n) / 2
# Calculate sum of ranks for each list.
let sra = sum(a.mapIt(if it: art else: arf))
let srb = sum(b.mapIt(if it: art else: arf))
let src = sum(c.mapIt(if it: art else: arf))
# Calculate H.
result = 12 / (n * (n + 1)) * (sra * sra / a.len.toFloat + srb * srb / b.len.toFloat +
src * src / c.len.toFloat) - 3 * float(n + 1)
proc runStudy(numSubjects = 1000; duration = 180; interval= 30) =
## Run the study using the population of size "numSubjects" for "duration" days.
var population = newSeqWith(numSubjects, Subject())
var unt, untCovid, irr, irrCovid, reg, regCovid = 0
echo &"Total subjects: {num_subjects}"
for day in 1..duration:
for subj in population.mitems:
subj.update()
if day mod interval == 0:
echo &"\nDay {day}:"
unt = population.countIt(it.category == Untreated)
untCovid = population.countIt(it.category == Untreated and it.hadCovid)
echo &"Untreated: N = {unt}, with infection = {untCovid}"
irr = population.countIt(it.category == IRREGULAR)
irrCovid = population.countIt(it.category == IRREGULAR and it.hadCovid)
echo &"Irregular Use: N = {irr}, with infection = {irrCovid}"
reg = population.countIt(it.category == REGULAR)
regCovid = population.countIt(it.category == REGULAR and it.hadCovid)
echo &"Regular Use: N = {reg}, with infection = {reg_covid}"
if day == duration div 2:
echo "\nAt midpoint, infection case percentages are:"
echo " Untreated : ", 100 * untCovid / unt
echo " Irregulars: ", 100 * irrCovid / irr
echo " Regulars : ", 100 * regCovid / reg
echo "\nAt study end, infection case percentages are:"
echo &" Untreated : {100 * untCovid / unt} of group size of {unt}"
echo &" Irregulars: {100 * irrCovid / irr} of group size of {irr}"
echo &" Regulars : {100 * regCovid / reg} of group size of {reg}"
let untreated = collect:
for s in population:
if s.category == Untreated:
s.hadCovid
let irregular = collect:
for s in population:
if s.category == Irregular:
s.hadCovid
let regular = collect:
for s in population:
if s.category == Regular:
s.hadCovid
echo "\nFinal statistics: ", kruskalWallis(untreated, irregular, regular)
randomize()
runStudy(10_000)
</syntaxhighlight>
{{out}}
<pre>Total subjects: 10000
Day 30:
Untreated: N = 8619, with infection = 283
Irregular Use: N = 1379, with infection = 27
Regular Use: N = 2, with infection = 0
Day 60:
Untreated: N = 7480, with infection = 498
Irregular Use: N = 2362, with infection = 77
Regular Use: N = 158, with infection = 3
Day 90:
Untreated: N = 6479, with infection = 712
Irregular Use: N = 2440, with infection = 151
Regular Use: N = 1081, with infection = 24
At midpoint, infection case percentages are:
Untreated : 10.98935020836549
Irregulars: 6.188524590163935
Regulars : 2.220166512488436
Day 120:
Untreated: N = 5655, with infection = 873
Irregular Use: N = 2199, with infection = 210
Regular Use: N = 2146, with infection = 70
Day 150:
Untreated: N = 4998, with infection = 1007
Irregular Use: N = 1910, with infection = 257
Regular Use: N = 3092, with infection = 135
Day 180:
Untreated: N = 4479, with infection = 1122
Irregular Use: N = 1575, with infection = 302
Regular Use: N = 3946, with infection = 232
At study end, infection case percentages are:
Untreated : 25.05023442732753 of group size of 4479
Irregulars: 19.17460317460317 of group size of 1575
Regulars : 5.879371515458693 of group size of 3946
Final statistics: 235.1089104422972
</pre>
Line 203 ⟶ 350:
{{trans|Raku}}
<syntaxhighlight lang="perl" line>use v5.036;
use List::AllUtils <
use enum qw<False True UNTREATED REGULAR IRREGULAR>;
use constant DOSE_FOR_REGULAR => 100;
Line 243 ⟶ 390:
}
my $H =
12/($n*($n+1)) * $H - 3 * ($n + 1)
}
Line 258 ⟶ 405:
@sirr = @hadcovid[ indexes { $_ == IRREGULAR } @category];
@sreg = @hadcovid[ indexes { $_ == REGULAR } @category];
( $unt, $irr, $reg) = (scalar(@sunt), scalar(@sirr), scalar(@sreg)
($hunt,$hirr,$hreg) = (
}
Line 702 ⟶ 849:
{{trans|Python}}
{{libheader|Wren-math}}
<syntaxhighlight lang="
import "./math" for Nums
| |||