Transcript DB16. Data Definitions HSV2: 1=, 0=not

Basic Conventions1
In a binary 0 or 1, 0 implies absence of the named
event and 1 implies its occurrence, e.g., a diagnosis,
a positive lab test, taking a medication, etc.
Age is in years unless otherwise indicated
Pt: patient; ptid: patient identification number
DX: diagnosis
TX: treatment
RT or XRT: radiation therapy
Ca: cancer, as CaP: cancer of the prostate
HIV: human immunovirus
AIDS: acquired immune deficiency syndrome
STD: sexually transmitted disease
HSV2: herpes simplex virus 2; genital herpes.
Basic Conventions2
CD4: Cell count of a type of T-cell that helps oppose
HIV. The higher the count, the healthier is the patient.
VL: Viral load; plasma level of viral RNA; a measure
of severity of viral disease. Log(VL) is usually used
because of strong right skew.
Alt, ast: alanine, aspartate aminotransferase;
measures liver health.
Cholesterol, triglycerides: measures heart health.
DB16. HSV2 in HIV Patients-1
• Background: HIV patients often have
comorbidities, one of which is the STD
HSV2.
• Q: Do people with/without HSV2 have
different CD4 counts and/or viral load
(VL)? (Stratify by HIV meds, which
artificially lowers VL and raise CD4.)
DB16. Data Definitions
HSV2: 1=infected, 0=not. Sex: 1=m, 0=f.
Yrshiv: no. of years pt has had HIV. Age:
years. Race: 1=EuroAmer, 2=AfroAmer,
3=HispAmer, 4=other. Cd4: CD4 count.
Vl1,vl2, vl3: 3 viral load assays. Control:
1=detectable VL, 0=not. Acyclovir: 1=on
med for HSV2, 0=not. Hivmeds: 1=on med
for HIV, 0=not. Sample of data format:
id
hsv2 sex yrshiv age race cd4 vl1
vl2
vl3
control acyclovir hivmeds
156
0 1
2 30
2 235 75000 14331 75000
0
0
0
413
0 1
9 42
4 709
0
0
0
1
0
1
DB17. Prostate Cancer in HIV Patients-1
• Background: HIV patients who acquire
prostate cancer (CaP) may follow a
different disease pattern than non-HIV
patients.
• Q: What characteristics are different in HIV
patients with and without CaP?
• Q: What variables predict CaP in HIV
patients?
DB17. Data Definitions
Ethn: 1=EuroAmer, 2=AfroAmer, 3=Hisp
Amer, 4=other. Hivyrs: years since HIV dx.
%cd4: %CD4 cells of all T cells. Lowcd4:
CD4 nadir. Haart: On HAART treatment or
not. Dxcap: Prostate cancer diagnosed or
not. Dre: Digital rectal exam positive or
not. On-testost: Being treated with testosterone or not. (Testosterone increases risk
of CaP.) Sample of data format:
Age ethn hivyrs cd4 %cd4 vl lowcd4 haart dxcap psa dre on-testost
36 1
9 954 26 52556 310
0
0
0.4 0
0
41 3
10 710 56 75000 538
0
0
0.7 1
0
DB18. Group A Strep Pneumonia Outbreak
• Background: A sudden outbreak of Group
A Streptoccum pneumoniae occurred at
the Marine Corps Recruit Depot in San
Diego, serious enough to have had special
coverage by CNN.
• Q: Did the Grp A strep pneumonia evince
any unique characteristics, i.e., different
from other pneumonias?
DB18. Data Definitions
Grpa: Group A Strep=1, other pneumonias=0.
Multilobe: x-ray shows more than one lung
lobe infected=1, single lobe=0.
Dayshosp: no. of days spent in hospital.
Remaining variables (lab test results, vital
signs, and symptoms) are obvious.
DB19. AIDS Mortality
• Background: Of a sample of 1224 AIDS
patients who died, time to death varied
with several influencing effects and
potential co-causes of death.
• Q: What variables affect the survival time
• of AIDS patients? How do they affect it?
• Q: Can survival time be predicted by
certain variables? If so, which?
DB19. Data Definitions
Serodate: Date of test showing pre-infection exists.
Cod: Cause of death. 1, 2: primary, secondary.
Pcp: pneumocystis carinii pneumonia.
Oi: Opportunistic infection.
Cmv: Cytomegalovirus.
Mac: Mycobacterium avium complex.
Cd4_200: CD4<200.
Waltreed_staging: staging (severity) of infection used
at Walter Reed Army Medical Center.
Gangacyclovir, fluconazole, septra, pentamidine,
azithro, dapsone: types of antibiotic medicine.
Hepbc, hepbs: hepatitis B core, surface antigen type.
Hepc: hepatitis C.
DB20. Erectile Dysfunction in
HIV Patients
• Background: Erectile dysfunction (ED) in
otherwise healthy adults has received
considerable attention, but little is known
about its pattern in HIV patients.
• Q: What is its rate of occurrence?
• Q: What HIV characteristics are different
between patients with and without ED?
DB20. Data Definitions
Ed_or_not: 1: has ED; 0: does not have ED.
Lowest cd4, current cd4: CD4 counts.
Current vl, highest vl: Viral Load counts.
Yrs_of_hiv: Years patient has had HIV.
Age: Patient’s years of age.
Haart_1: Pt has had HAART treatment: 1;
has not had HAART treatment: 0
DB21. HIV Patients with Syphilis
• Background: Co-morbidities are a serious
problem with HIV patients. Syphilis is a
severe one, but little is known about its
effect on the health of these patients.
• Q: What health characteristics of HIV
patients change upon contracting
syphilis?
DB21. Data Definitions
• Race: 0=EuroAmer, 1=AfrAmer,
2=HispAmer, 3=AsianAmer.
• Before: before contracting syphilis.
• After: after contracting.
• Hepbc, hepbs: hepatitis B, core and surface
antigen type, respectively.
• Hepc: hepatitis C.
• Hsv: herpes simplex virus 2 (genital).
• Everonhaart: has been treated with HAART
med cocktail.
DB22. Pediatric Snake Bites
• Background: Rattlesnake bite data from
San Diego Poison Center on 204 children
• Q: Is location of bite affected by age, sex,
and/or season?
DB22. Data Definitions
• Age: years. Sex: male = 1, female = 2.
• Agegroup: 2 = 13 (toddler), 7 = 49 (child),
11 = 1012 (tween), 15 = 1318 (teen).
• Location: 2/3 = hand/arm, 4 = face,
5/6 = leg/foot, 7 = unknown.
• Locbinary: 2 = hand/arm, 5 = leg/foot.
• Month of Bite: 1 = Jan, 2 = Feb, etc.
• Season: 1 = winter (months 1, 2, 3),
2 = spring (months 4, 5, 6), etc.
DB23. Adult Snake Bites
• Background: Rattlesnake bite data from
San Diego Poison Center on 625 adults
• Q: Is location of bite affected by age,
sex, and/or season?
• Q: Are adult findings different from
those for children (DB22)?
DB23. Data Definitions
• Age: years. Sex: male = 1, female = 2.
• Agetrinary: 25 = <35, 45 = 35-55, 65 = >55.
• Location: 2/3 = hand/arm, 4 = face,
5/6 = leg/foot, 7 = unknown.
• Locbinary: 2 = hand/arm, 5 = leg/foot.
• Month: 1 = Jan, 2 = Feb, etc.
• Season: 1 = winter (Jan/Feb/Mar), 2 = spring
(Apr/May/Jun), etc.
DB24. Weight Estimation in the ER
• Background: Many drug dose levels
depend on the patient’s weight. In the
ER, patients are not always lucid. This
DB has true measures, various
estimates, and various covariates that
could be used in an estimation equation.
• Q: What is the best way to estimate a
patient’s weight? Is this estimate
reliable?
DB24. Data Definitions
•
•
•
•
•
•
•
•
Age: years. Sex: 1 = female, 2 = male.
Ht: height (cm). Wt: weight (kg).
Est: estimate.
Nok: next of kin. Phys: physician.
Inv: investigator.
Circ: circumference. Abd: abdomen.
1 & 2: repeated readings by 2 physicians.
Tibia length: standardized length of tibia.
DB25. Snoring as a Sleep Problem
• Background: Snoring is thought of as an
adjunct to sleep problems, not a cause.
Accepted causes are apnea, hypopnea,
obesity, age, et al.
• Q: Do our data show that snoring is a
cause of sleep problems?
• Q: What other causes do the data show?
DB25. Data Definitions
•
•
•
•
•
•
•
•
•
•
•
Epworth: sleepiness scale. ESS & FOSQ: range 024
Fosqtotal: Functional Outcome of Snoring Quotient
Age: years. Sex: 1 = female, 2 = male
Bmi: body mass index = 703*weight (lb)/height2 (in)
Ahi: apnea-hypopnea index—episodes per hour av
Dur: duration (hrs) of snoring > 50 dB
Max: maximum dB level of snoring
Avdb: mean dB level of snoring when > 50 dB
Percent: % of sleep time when mean dB > 50
Snoring, obesity,…, stroke: in pt’s record (1)
…byspouse: bed partner complains
DB26. Vulvar Cancer Survival
• Background: DB has 45 vulvar cancer
patients with numerous putative
correlates, e.g., age, overweight, tobacco
use, immunocompromise, diabetes,….
• Q: What variables relate to cancer stage?
• Q: What variables relate to number of
positive lymph nodes?
• Q: What variables predict survival?
DB26. Data Definitions1
•
•
•
•
•
•
•
•
•
•
•
Stage: Of Ca. 15 respectively: stage 1a, 1b, 2, 3, 4.
Nodes: Affected (1) or not (0). Dead: yes (1) or no (0).
Years: Diagnosis to death if dead or last f-u if not.
Gravididity: # pregnancies. Parity: # births.
BMI: body mass index = 703xlbs/in2.
Tobacco: 1 = current, 2 = quit, 3 = never.
Warts: No. of warts.
Degreevin: Severity of VIN infection on 15 scale.
Sxlength: Duration (years) of symptoms.
Los: Length of surgery (hours).
Ebl: Estimated blood loss (ml) during surgery.
DB26. Data Definitions2
• Histology: Squamous cells (1) or non-squamous (2).
• Nodestotal: Number of affected lymph nodes.
• Present (1) or absent (0) conditions:
• Hpv (human papillomavirus),
• condyloma (wart-like growth on vulva),
• vin (a related virus),
• dm (diabetes),
• steroids (using or not),
• imcomp (immunosupressed),
• margins (Ca spread outside locality),
• xrt (radiotherapy or not),
• chemo (chemotherapy or not).
DB27. Artificial Hip Placement
• Background: Artificial hips must be
carefully positioned to be effective.
Operated and non-operated leg lengths
and hip angles are compared for 69
patients, 38 with x-ray guidance and 31
without, all read twice for reliability.
• Q: Does x-ray during placement increase
accuracy?
• Q: Are the readings reliable between and
within readers?
DB27. Data Definitions
•
•
•
•
•
•
•
•
Obsrvr. Reader 1 or 2
X-ray. 1: x-ray used, 0: not used
Pre, preop, postop. Pre- or post-operative
Op, non. Operated or non-operated hip
1, 2, av. 1st, 2nd, mean rdg. by same reader
Lld. Leg length difference between readings
Abdang. Abduction angle (from vertical)
Distances in centimeters, angles in degrees
DB28. Does Handling Worsen Injuries
• Background: Badly injured (wounded)
patients trucked or airlifted to hospital
are roughly handled and become cold.
Ten pigs are injured, cooled, and roughly
handled, and their conditions measured.
• Q: What is the effect of cooling and/or
handling? When it is severe enough to
risk death?
DB28. Data Definitions
The pigs are cooled progressively until
their core temperatures reach prescribed
levels, then are jostled and dropped from a
few-inch height. Heart rate (HR), mean
arterial pressure (MAP), pH, and CO2 are
measured initially and at each level, along
with incremental number of minutes to
reach that level. Levels were 38º, 35º, 32º,
29º, 26º, 23º, and 20º. Death or not is
recorded, with final temperature.
DB29. Compare OR Thermometer Types
• Background: In the OR, different types of
temperature measurement used were (1)
esophageal, (2) tympanic, and temporal
arterial scanner (TAS) on (3) forehead
and (4) forehead plus behind ear.
• Q: Do they measure the same? If not, can
one be made equivalent to another by a
linear transform?
• Q: Is each internally consistent?
DB29. Data Definitions
Five temperature measurements (temp) are
made 10 minutes apart on each of 48
patients by four types of thermometer:
eso: esophageal
tym: tympanic
for: TAS on forehead
ear: TAS on forehead + ear
Example. fortemp2: the 2nd reading of the
TAS on the forehead
DB30. Laryngectomy Survival
• Background: Three modes of therapy for
larynx cancer are radiation therapy (RT),
larynx conservation surgery (CS), or a
combination. When therapy fails, a salvage
total laryngectomy (STL) may be tried.
• Q: What is survival rate through time?
• Q: Does survival after laryngectomy vary
with type of initial therapy?
DB30. Data Definitions
Retrospective study, giving
survival outcomes of STL
(died: 1 = died, 0 = not) and
years to death or last follow up
for
Grp 1: 53 initial RT failures
Grp 2: 6 initial CS failures
Grp 3: 5 initial RT+CS failures
DB31. Positioning Patients in H&N RT
• Background: Patients are held in a helmetmask for repeated H&N (head & neck) RT to
localize RT focus on the tumor and avoid
adjacent tissue. Independent exact
measurement of position’s deviation from 0.
• Q: Is the RT focus accurate on average?
• Q: Does the RT focus change in a patient
over repeated treatments? If so, in what
dimensions?
• Q: Does therapist or machine affect
accuracy?
DB31. Data Definitions
•
•
•
•
•
•
•
•
•
•
13 pts with 20-30 treatments each: 320 obs.
Sessnr: session number. Nrtx: no. treatmts.
x, y, z: position in 3 respective dimensions.
Octant: position of error in 1 of 8 directions.
Therapist: ID of person administering RT.
Machine: ID of radiation machine used.
Firstlast: marks 1st, last session per patient.
Wtdif: Weight loss (lbs) during therapy.
Sex: 1 = m. Age: years.
Dose: amount of radiation.
DB32. Anesthetic Type v. Recovery Time
• Background: There is a variety of
anesthetics, which vary with dosage,
with type of procedure, and with patient
characteristics.
• Q: Does anesthetic type influence time to
start procedure, length of procedure,
and/or time to recovery? Do dose, age,
and/or weight affect these variables?
DB32. Data Definitions
• Anestype: primary anesthetic type (some
have others added, not dominating effect).
1: ketamine IV, 2: ketamine IM, 3: a narcotic
(versed, fentanyl, morphine), 4: propofol.
• Dose: standardized dose, i.e., scaled to
same level of intended anesthesia.
• Age: years.
• Wt: weight kg. (Most adults defaulted to 70.)
• Procedure time, recovery time, discharge
(dc) times in minutes.
DB33. Reliability in Reading X-Rays
• Background: Interpretation of an x-ray
may profoundly affect a patient.
• Q: Does a radiologist read the same
result consistently?
• Q: Do different radiologists read the
same result?
• Q: How important are training and
experience in reading x-rays?
DB33. Data Definitions
• Three radiologists (G, C, and S) judge
disease severity shown in radiographs from
100 cancer patients.
• Each reads each x-ray 3 times: once, wait a
month, shuffle order, reread, wait a month,
shuffle order, read a third time.
• Each reading is a judgment of cancer stage,
i.e., how far cancer has progressed, on a 0
to 5 scale.
DB34. Disruptive Back Pain in Pilots
• Background: Helicopter pilots suffer a
disproportionate amount of back pain
which affects their safety and efficiency.
• Q: What factors are associated with pilot
back pain? (Type of aircraft? Length of
flight? Frequency of duty? Height?
Obesity? Use of medications? Presence
of neurologic symptoms? Etc.)
DB34. Data Definitions
• Presence (1) or absence(0):
Presence of back pain, affects flight, seen
by physician, taking medications,
neurologic symptoms.
• Age (years), height (in.), weight (lbs.).
• Type of aircraft: 1 = AH-1W, 2 = UH-1N.
• Days per week: days of flight per week
regularly scheduled.
DB35. Effect of Fasting on
Cholesterol Measurement
• Background: Conventional wisdom says
that cholesterol measurement is not
reliable unless taken after fasting.
• Q: Are fasting and non-fasting
cholesterol readings different?
• Q: Are they internally consistent?
• Q: Can fasting cholesterol be estimated
from non-fasting cholesterol?
DB35. Data Definitions
• Sex1f: Female (1), male (2). Age: years.
• Suffix fast implies measurement taken after
fasting, eat implies without fasting.
• Tot: Total cholesterol.
• Ldl: Low density lipoprotein level.
• Hdl: High density lipoprotein level.
• Tg: Triglyceride level.
• Glu: Glucose level.