Skip to main content

Table 1 Candidate predictors of bacteriuria measured (a) once at admission (constant throughout one hospital stay; time independent) and (b) multiple times throughout a patient’s hospital stay (time-dependent)

From: Development of risk prediction models to predict urine culture growth for adults with suspected urinary tract infection in the emergency department: protocol for an electronic health record study from a single UK university hospital

Candidate predictor

Definition

Units/categories

(a) Measured at admission

Demographic

 Age

Recorded age at hospital admission in 10-year age bands (continuous age is unavailable due to privacy regulations)

18–24, 25–34,…, 95–104

 Sex

As recorded in the admission notes

Male/female

 Social deprivation

Index of Multiple Deprivation (IMD) 2015 quintile

Deciles (1–10)

 Ethnicity

As recorded in the admission notes; collapsed into 5 major categories

Asian, Black, Mixed, other, White

Co-morbidity

 Charlson Co-morbidity Index

Numeric comorbidity score based on the presence of relevant ICD-10 codes in the entire hospital record*

Count (1–33)

 Underlying renal disease

Presence of a relevant ICD-10 code in the previous 5 years*

Yes/no

 Underlying urological condition

Presence of a relevant ICD-10 code in the previous 5 years*

Yes/no

 Renal or urological surgery

Presence of a relevant OPCS code in the previous 5 years*

Yes/no

 Immunosuppression

Presence of a relevant ICD-10 code in the prior year*

Yes/no

 Cancer

Presence of a relevant ICD-10 code in the prior year*

Yes/no

Previous healthcare contact

  

 Discharge from hospital in prior 7 days

Most recent discharge date from QEHB within 7 days of index attendance date

Yes/no

 Number of previous admissions

Number of hospital spells at QEHB in the prior year

Count (≥ 0)

 Number of days spent in hospital

Number of days spent as an inpatient at QEHB in the prior year

Count (≥ 0)

 Number of previous ED attendances

Number of ED attendances at QEHB in the prior year

Count (≥ 0)

Factors predisposing to UTI

  

 Previous admission for UTI

Admission to QEHB with an ICD-10 code of UTI on discharge in the prior year*

Yes/no

 Previous ED attendance for UTI

ED attendance at QEHB with ED diagnosis of lower UTI, pyelonephritis or urosepsis in the prior year

Yes/no

 Number of previous admissions for UTI

Number of hospital spells at QEHB with an ICD-10 code of UTI on discharge in the prior 2 years*

Count (≥ 0)

Number of previous ED attendances for UTI

Number of ED attendances at QEHB with ED diagnosis of lower UTI, pyelonephritis or urosepsis in the prior year*

Count (≥ 0)

 Previous urine culture

Urine sample submitted at QEHB for microbiological diagnosis in prior year

Yes/no

 Previous bacteriuria

Urinary pathogen identified at QEHB from blood or urine in prior year

Yes/no

 Previous resistant pathogen

Drug-resistant pathogen identified at QEHB from blood or urine in prior year

Yes/no

 Prior antibiotic consumption

Total antibiotic consumption in QEHB in prior year

Defined daily doses (DDDs) (≥ 0) [18]

Characteristics of the admission

 Admitted from care home

As recorded

Yes/no

 Month of admission

As recorded

January,…, December

 Day of year of admission

As recorded

Count (1–366)

 Day of week of admission

As recorded

Monday,…, Sunday

Investigations in the ED

 Suspected diagnosis in the ED

ED impression of clinical syndrome as recorded by the ED clinician

Lower UTI, pyelonephritis, urosepsis

 Positive urinalysis

Presence of leucocytes and/or nitrates in urinalysis

Yes/no

Urinalysis

 

 Leucocytes

As recorded by the clinician (dipstick test)

Positive/negative

 Nitrates

As recorded by the clinician (dipstick test)

Positive/negative

 White blood cells

As recorded by the laboratory (flow cytometry)

Count/μL

 Red blood cells

As recorded by the laboratory (flow cytometry)

Count/μL

 Epithelial cells

As recorded by the laboratory (flow cytometry)

Count/μL

 Small round cells

As recorded by the laboratory (flow cytometry)

Count/μL

 Bacteria

As recorded by the laboratory (flow cytometry)

Count/μL

 Yeast

As recorded by the laboratory (flow cytometry)

Count/μL

 Conductivity

As recorded by the laboratory (flow cytometry)

mS/cm

 Casts

As recorded by the laboratory (flow cytometry)

Count/μL

 Crystals

As recorded by the laboratory (flow cytometry)

Count/μL

(b) Measured multiple times throughout hospital stay

Clinical observations

 Heart rate

As recorded

Beats per minute

 Respiratory rate

As recorded

Breaths per minute

 Body temperature

As recorded

 Oxygen saturation

As recorded

Percent

 Systolic blood pressure

As recorded

mmHg

 AVPU

As recorded

Alert, verbal, pain, unresponsive

 SEWS

Standardised Early Warning Score as recorded or calculated based on heart rate, respiratory rate, body temperature, oxygen saturation and AVPU

Count (0–18)

Clinical investigations

 White cell count (blood)

As recorded

103/mL

 C-reactive protein

As recorded

mg/L

 Creatinine

As recorded

μmol/L

 Acute kidney injury score

Defined as the change in serum creatinine compared to an approximate baseline measure (i.e. average creatinine in previous 6 months)

Stage 0 (1.0–1.5 × baseline), stage 1 (1.5–1.9), stage 2 (2.0–2.9) stage 3 (≥ 3.0)

 Alkaline phosphatase

As recorded

IU/L

 Bilirubin

As recorded

μmol/L

 Platelets

As recorded

109/L

Antibiotic treatment

 Antibiotic treatment

Recorded administration of any systemic antibiotic (British National Formulary chapter 5.1.)

Yes/no

 Broad-spectrum antibiotic

Recorded administration of any of the following antibiotics: co-amoxiclav, piperacillin-tazobactam, carbapenems, cephalosporins (except 1st generation), quinolones, colistin, fosfomycin, aminoglycosides

Broad-spectrum, narrow-spectrum, none

 Route of administration

Recorded route of administration, giving precedence to intravenous (IV) antibiotics (i.e. if multiple antibiotics are prescribed with ≥ 1 IV, treatment is classified as IV)

IV, oral, none

 Dosage

As recorded

DDDs (≥ 0) [18]

  1. *Detailed code lists available in the appendix
  2. For each time-dependent variable, we will also consider the change in value compared to the last observed measurement
  3. Excluding anti-tuberculosis and anti-leprosy medication