Skip to main content

Table 2 Identification criteria utilized in the published studies. Part B: Studies with high risk of bias (n = 23, in chronological order)

From: Algorithms to identify COPD in health systems with and without access to ICD coding: a systematic review

Authors, publication year, country Dataset used Study population / COPD population Age limitation ICD codes Hospitalization Ambulatory visit Physician claims (ambulatory) Pharmacotherapy Spirometry Oxygen therapy COPD-related surgical procedure Algorithm Risk of bias (low or high)
Hansell et al. (2003)
United Kingdom [37]
- Office for National Statistics
- Hospital Episode Statistics
- General Practice Research Db
- Health Survey for England 1995
SP: NA
COPD-P: NA
  X X X X X     COPD: ICD-9: 490–492, 494–496.
COPD symptoms: cough or phlegm for at least 3 months during the winter
high
(no validation of algorithm)
Wilchesky et al. (2004)
Canada [38]
- Quebec universal medical insurance register
- Medical services claims data
SP: NA
COPD-P: 14,980
X X    X      COPD: 490–490.9, 494–494.9, 496–496.9
Diagnostic criteria:
1. Years ≥66.
2. Two visits or more to the MOXXI physicians in the year.
high
(validation against diagnostic codes from medical charts: low sensitivity 45.97 [43.9,48.0] but higher specificity 88.42 [87.9,89.0]
Lacasse et al. (2005)
Canada [39]
- Quebec universal medical insurance register SP: 2,487,605
COPD-P: 176,313
X X X X X X     Years ≥65.
COPD: IDC-9: 491, 492 and 496.
Possible COPD: Years ≥65 or older, always registered as COPD (never as asthmatics), appeared three times in the database and who field prescription for ipratropium bromide or beta2-agonist
Probable COPD: all above plus internist or pulmonologist COPD diagnosis.
high
(no validation against reference standard; sensitivity and specificity were not determined)
Mapel et al. (2006)
United States [11]
- Lovelace Health Plan, a health maintenance organization serving New Mexico SP: 41,428
COPD-P: 2129
X X X X   X     Years ≥40.
Any patient with one or more claims records with COPD diagnosis (491, 492, 496).
Excluded: ICD-9140–208, 494, 405, 500–519, 173, 174, 185.
high
(algorithm: low sensitivity 60.5%, specificity 82.1%.
Reference standard: COPD diagnosis abstracted from medical record, based on ICD codes.)
Akazawa et al. (2008)
United States [40]
- United Healthcare claims data (medical and pharmaceutical) SP: 81,322
COPD-P: 28,968
X X X   X X     Years ≥40.
1. Inpatient hospital or emergency room bill with a diagnosis with ICD code: 491, 492, 496; or
2. Physician claims with a COPD diagnosis and a second COPD-related medical claim with a separate date; or
3. Physician claims with a COPD diagnosis with pharmacy claim for certain medication.
high
(no validation of algorithm)
Heins-Nesvold et al. (2008)
United States [41]
- Managed care administrative Db (Medical and pharmacy administrative claims data, Midwest health)
- Mailed survey to COPD patients in Minnesota
SP: NA
COPD-P: 7782
X X X X X X X X   Cases were identified through enrolment, pharmacy and medical files.
2. Years ≥40
3. At least one claim with a COPD diagnosis
4. Claim with the COPD associated diagnosis initiated from a “medical” place of service.
high
(no validation of algorithm)
Mapel et al. (2010)
United States [42]
- Lovelace Health Plan SP: 10,904
COPD-P: 2707
X X X X   X     Years ≥40.
COPD: ICD-9 code 491, 492, 496.
Continuously enrolled for 2 years prior to index date, one inpatient or two outpatient claims with COPD-related ICD code and National drug COPD-related code.
high
(algorithm: specificity 70.5%, but low sensitivity 60.6%.
Reference standard: COPD diagnosis in a medical record)
Dalal et al. (2011)
United States [43]
- IMS Lifelink Db SP: NA
COPD-P: 9188
X      X     Years ≥40
PD maintenance medication pharmacy claim, specially inhaled corticosteroids, long-acting beta-agonists, anticholinergics, and/or fixed-dose combination regimes.
high
(no validated algorithm)
Mapel et al. (2011)
United States [44]
- US managed care administrative claims data (multiple health plans) Commercial insurance:
SP: 7,671,018 COPD-P: 42,565
Medicare insurance:
SP: 115,652 COPD-P: 8507
X X X X X     X 1. Years ≥40; one inpatient COPD hospitalization or emergency department visit (ICD 491, 492, 496); or
2. Years ≥40; two professional COPD claims (different service dates); or
3. Years ≥40; surgical procedure related to the COPD (lung volume reduction).
high
(no validated algorithm)
Dalal et al. (2012)
United States [45]
- Ingenix Impact National Benchmark Db SP: NA
COPD-P: 1936
X X     X     1. Years ≥40
2. Continuously enrolled
3. Received maintenance therapy: anticholinergic or fluticasone propionate/salmeterol combination within 1 month after an index event
COPD IDC-9 code: 491, 492 and 496
high
(no validated algorithm)
Make et al. (2012)
United States [46]
- PharMetrics Db including 12,4 million covered lives Commercial insurance:
SP: 7.671,018
COPD-P: 42,565
Medicare:
SP: 115,652
COPD-P: 8507
X X X X X     X Years ≥40.
Any of the following:
1. One inpatient COPD-related hospitalization or one emergency department visit (ICD-9: 491, 492, 496)
2. Two professional COPD claims (different service dates)
3. Surgical procedure related to the COPD listed on a facility or professional claim.
At least one filled prescription for drug during the study period was enough to consider the patient as to be taking medication. Maintenance COPD pharmacotherapy: LABA, SAAC, LAAC, theophylline and inhaled corticosteroids. SABA was considered symptomatic medications.
high
(no validated algorithm)
Gini et al. (2013)
Italy [47]
- Hospital discharge records
- Drug dispensing records
- Disease-specific exemption from co-payment to health care
- Inhabitant Registry
SP: 11,656
COPD-P: NA
  X X X   X     Identification of COPD patients was performed with use of:
Hospital discharge records (ICD codes: 490–492; 494, 496); drug dispensation records (ATC code); general physician data (ICD code: 490–492, 494, 496).
high
(no validation of algorithm)
Macaulay et al. (2013)
United States [48]
- Geisinger Health System (GHS) SP: NA
COPD-P: 2028
  X X X X X X    COPD ICD-9 codes: 491, 492 or 496 .
Results from at least one spirometer test
Reference standard: COPD diagnosis (using ICD-9 codes) and electronic health record results from at least one spirometry test.
high
Validation of COPD severity Code: low sensitivity, high specificity only for severe/very severe category)
Yawn et al. (2013)
United States [49]
MarketScan® Db:
1. Commercial Claims and Encounters
2. CMS Supplemental and Coordination of Benefits
SP: 1,669,546
COPD-P: 135,445
X X X X   X     Years ≥45.
COPD: ICD - 9 codes: 491, 492 or 496.
1. Admissions or emergency department visits or at least two COPD-related office visits with different service dates. A continuous enrolment of patients was required for the period of 1 year before the COPD diagnosis and at least 2 months after the COPD diagnosis date.
Excluded: with a history of ICS use or pneumonia in the 1 year baseline period; asthma, cystic fibrosis and lung cancer.
high
(no validated algorithm)
Dore et al. (2014)
United States [50]
- Normative Health Information Db (UnitedHealth Care) SP: NA
COPD-P: 225,079 LABA Users
X X     X     Years > 20.
COPD: ICD-9491.2, 492.8, 496.
- Top 3 variables predictive for COPD confirmation: 65 or older, inhaled anticholinergic drug and radiologic examination of the chest
- Claim for COPD Only: medications, prescriber specialty, diagnoses, spirometer procedure.
high
(Validation of algorithm against medical records – low sensitivity)
Erdem (2014)
United States [51]
- Chronic Conditions Public Use Files (Centers for Medicare and Medicaid Services) SP: NA
COPD-P: NA
  X X X       COPD: ICD-9 code, CPT-4 code or the HCPCS code. high
(no validated algorithm)
Vozoris et al. (2014)
Canada [52]
- Ontario Health Insurance Plan claims Db
- Canadian Institute for Health Information Discharge Abstract Db
- Ontario Mental Health Reporting System
- National Ambulatory Care Reporting System Db
- Same-Day Surgery Db
- Registered Persons Db
- Ontario Drug Benefit claims Db
SP: NA
COPD-P: 177,355
X X X X X      Years > 66.
COPD: At least three ambulatory claims for COPD within 2 years, or at least one COPD hospitalization
high
Algorithm: specificity 95.4%, low sensitivity 57.5%
Aldrich et al. (2015)
United States [53]
- Center for Medicare and Medicaid Services encounter SP: 26,927
COPD-P: 20,945
X X X X X      Years 40–79.
COPD diagnoses defined by using two previously published algorithms (Stein et al. 2012, Mapel et al. 2011).
1. Mapel: one or more COPD hospitalization or emergency department visit (ICD-9491, 492, 496) or at least two professional claims (different service dates)
2. Alternatively, a primary discharge COPD diagnosis (ICD-9491.21) throughout the same period of time following algorithm four defined by Stein et al.
high
Validity: low sensitivity 62% and positive predictive value of 80% for identified COPD.
Reference standard: COPD diagnosis in reviewed medical record
Vozoris et al. (2015)
Canada [54]
- Ontario Drug Benefit Db
- Ontario Health Insurance Plan Db
- Canadian Institute for Health Information Discharge Abstract Db
- National Ambulatory Care Reporting System Db
- Ontario Mental Health Reporting System
- Same-Day Surgery Db
- Ontario Cancer Registry
- Database of Ontario adults with physician-diagnosed congestive heart failure
- Registered Persons Db
Community dwelling:
107,109
Long-term care resident:
16,207
X X X X X      Years ≥66.
COPD diagnosis algorithm used three or more COPD ambulatory claims within a period of 2 years or at least one COPD hospitalization (specificity 95.4%, sensitivity 57.5%)
1. Three or more ambulatory claims for COPD within 2 year period or
2. One or more hospitalizations for COPD
3. Medication records.
high
Algorithm: specificity 95.4% [95% CI 92.6–97.4%]; low sensitivity 57.5% [95% CI 47.9–66.8%])
Laforest et al. (2016)
France [55]
- the Permanent Sample of Health Insurance Beneficiaries (EGB): a 1/97th random sample of the French National Claims Data beneficiaries (SNIIRAM) with individual linkage between primary (ambulatory) and secondary (hospital) care SP: 4237
COPD-P: 4237
X X X    X     years ≥45
1. COPD related hospitalization (ICD-10 codes J41, J42, J44 and J96.1.. The J96.0 was accepted as primary diagnosis only if J43 or J44 were present)
2. Long-term disease status for COPD (ICD-10 codes J41, J42, J44 and J96.1)
3. Bronchodilator drugs (LABA, SABA, LAMA, SAMA, xanthines, and SAMA/SABA fixed combinations.
high
(no validation of algorithm against clinical reference standard)
Price et al. (2016)
United States [56]
Clinformatics™ Data Mart retrospective claims database:
- include medical claims (primary and secondary care),
- pharmacy claims and
- laboratory test results
SP: 93,980
COPD-P: 6687
X X    X X     years 4–64
1. Diagnosis of COPD and/or
2. Exercise induced bronchoconstriction recorded at any time and
3. At least one prescription for albuterol
high
(no validation of algorithm against clinical reference standard)
Raymakers et al. (2017)
Canada [57]
- PharmaNet prescription data
- Discharge Abstract Database
- British Columbia Vital Statistics Deaths
- The regional health authority and census neighborhood income data
- Physician billing data from the provincially administered universal insurance program
SP: 39,678
COPD-P: 41,602
X      X     years ≥50
1. Three or more prescriptions (anticholinergic or a short-acting beta agonist) in a 12-month period
Index date: the date of receipt of the first prescription
high
(no validation of algorithm against clinical reference standard)
Turner et al. (2018)
United States [58]
- HealthCore Integrated Research Database
- Medical records
SP: 2,219,034
COPD-P: 17,156
X X    X X X    years ≥40
1. ≥2 COPD diagnoses (ICD-9 CM codes 491, 492, 496),
2. ≥2 COPD-related procedures,
3. ≥3 Generic Product Identifier (COPD medication prescription fills) and
4. ≥2 Current Procedural Terminology codes for spirometry tests
high
medical record review: COPD confirmation by persistent airflow obstruction FEV1/FVC < 0.70 at symptom baseline; but missing data constrained COPD identification
  1. The next-to-last column on the right gives the identification criteria based on the statements contained in the publication
  2. SP Study population, COPD COPD-P population, Db Database, NA Not available; see also list of abbreviations