Prevalence of Angina Among Primary Care Patients With Coronary Artery Disease

Key Points Question What is the prevalence of angina among stable US outpatients with coronary artery disease (CAD)? Findings In a survey study of 1612 outpatient primary care patients with CAD in a large US integrated primary care network, 21.2% of surveyed patients reported experiencing angina at least once per month (daily or weekly, 12.5%; monthly, 8.7%). After multivariable adjustment, speaking a language other than Spanish or English, Black race, smoking, atrial fibrillation, and chronic obstructive pulmonary disease were associated with increased angina frequency. Meaning These findings suggest that angina is prevalent among US outpatients with CAD; proactive angina assessment in outpatient settings may identify patients with suboptimally controlled angina and may be associated with improved treatment and outcomes.


PCOI Coronary Artery Disease Criteria:
a. Inclusion -Adult patients (18 and older) with the following: 1) One CPT-4 procedure code or procedure list term for CAD intervention OR 2) Two outpatient visits with 2 problem list terms or 1 ICD billing diagnosis plus 1 problem list term for CAD b. Exclusion-None 2. Data Sources: (look back time frames) a. Pre-EPIC Implementation: 1) Billing Procedure -CPT (All history) 2) LMR problem listkey word (All history) 3) LMR Procedure listkey word (3 years) 4) Oncall problem and procedure listkey word (3 years) 5) Billing diagnosis -ICD9/ICD10 (3 years) 410.00 Acute myocardial infarction of anterolateral wall, episode of care unspecified 410.01 Acute myocardial infarction of anterolateral wall, initial episode of care 410.02 Acute myocardial infarction of anterolateral wall, subsequent episode of care 410.10 Acute myocardial infarction of other anterior wall, episode of care unspecified 410.11 Acute myocardial infarction of other anterior wall, initial episode of care  (s) three venous grafts; (List separately in addition to code for primary procedure) 33521: Coronary artery bypass, using venous graft(s) and arterial graft(s) four venous grafts; (List separately in addition to code for primary procedure) 33522: Coronary artery bypass, using venous graft(s) and arterial graft(s) five venous grafts; (List separately in addition to code for primary procedure) 33523: Coronary artery bypass, using venous graft(s) and arterial graft(s) six or more venous grafts; (List separately in addition to code for primary procedure) 33530: Reoperation, coronary artery bypass procedure or valve procedure, more than one month after original operation (List separately in addition to code for primary procedure) 33533: Coronary artery bypass, using arterial graft(s); single arterial graft 33534: Coronary artery bypass, using arterial graft(s); two arterial grafts 33535: Coronary artery bypass, using arterial graft(s); three arterial grafts 33536: Coronary artery bypass, using arterial graft(s); four or more arterial grafts 92980: Transcatheter placement of an intracoronary stent(s), percutaneous, with or without other therapeutic intervention, any method; single vessel 92981: Transcatheter placement of an intracoronary stent(s), percutaneous, with or without other therapeutic intervention, any method; each additional vessel (List separately in addition to code for primary procedure) 1) Inclusion-Adult patients (18 and older) with a diagnosis of CAD within the prior 3 years. Chart review diagnosis is based upon a problem list that includes CAD or related terms (see above list of terms) and a procedure/s that would be appropriate for treating CAD (see procedure code list above). If a patient has a problem list that includes CAD or related terms but no CAD procedure, then the patient was considered to have CAD if the details of the CAD problem list indicated that the patient was considered to have definitive CAD from the provider's perspective. If the patient did not have a problem list that included CAD but had a procedure related to CAD done, it was presumed that the patient has CAD. Other qualifying criteria include PCP or Cardiology notes that discuss CAD status, history or treatment. Individuals treated for CAD in the past, but with stable disease are still considered as having CAD. 2) Exclusion-Patients who do not have a diagnosis of CAD based upon problem list review, or have not had a procedure/s to treat CAD noted.

1) Development:
The process started with a review of our past algorithm, Vascular Disease Documentation-original algorithm2008.doc. We used the original algorithm, which only searched the LMR/Oncall problem/procedure lists to generate an initial CAD population using the 2009 linkage cohort. From this group, a random sample was selected for review. This review identified frequently used search terms for problems and procedures. It also provided a list of non-specific terms that did not define CAD with sufficient specificity (i.e. "stent"). After multiple reviews of this exhaustive list of potential items, a final list of terms that were considered to identify CAD (inclusion terms) and non-specific items (exclusion terms) was agreed upon. A review of original CPT and ICD9 codes compared to updated information resulted in a new list of the most current codes available. To determine whether ICD-9 diagnosis billing codes should be used in addition to EHR problems, a comparison of diagnoses based upon problem lists vs. problem lists plus ICD-9 codes was performed. This review found that only 3% of patients without a problem list for CAD had an ICD-9 diagnosis code for CAD. To examine the potential value of these extra patients, a list of 20 randomly selected patients was pulled based on ICD9 codes for CAD from billing data. For the 20 patients with ICD9 codes for CAD, 7 (35%) were found not to have CAD. Given the small increase in cases and the low specificity of the ICD-9 diagnosis code in the absence of a CAD problem, ICD9 codes were not used in the final algorithm to identify patients with CAD. Using this final list of problems and procedure terms from the EHR and CPT codes from billing data, the algorithm was updated and MRNs identified as having CAD were pulled for review.
2) Validation: Using the new list of inclusion and exclusion terms and new codes, two lists were pulled from the 2009 PCOI linkage cohort: 1) a list of 15 patients from each practice with CAD based upon the new algorithm, and 2) a list of 10 patients per practice without CAD based upon the new algorithm. A chart review of these 325 patients (13 practices with 25 patients per practice) was performed to validate the algorithm's designation of coronary artery disease status or not. Overall, the results of the review for the search terms and CPT codes demonstrated excellent sensitivity and specificity (Sensitivity =96%, Specificity= 99%, PPV= 99%, and NPV= 95 %.)

3) 2014 Update:
In 2012-2013, our vascular disease algorithms were introduced into the Top Care population management tool. As more and more clinicians used Top Care, it was found that the sensitivity of our algorithm, especially PVD needed to be more sensitive. When our original algorithms were developed, we did not include ICD-9 billing or diagnosis codes. Looking to improve the sensitivity of all three algorithms, we added ICD-9 codes as well as any new problem or procedure terms to our search. As with our previous validations, we pulled 15 patients per practice with one or more of our vascular disease categories (CAD, CVD, PVD), and 10 patients per practice with no vascular disease. A blinded cohort of 936 patients was reviewed for validation. After validation we found that for CAD, the addition of the codes did not improve our sensitivity but did not decrease it. For both CVD and PVD, the sensitivity of the algorithm increased.

4) 2015 Update:
In October 2015, ICD-10 codes were introduced to replace ICD-9 for diagnosis billing. We updated our algorithms to reflect this change.
No formal validation was done at this time. We did use the new version of the algorithms to compare the numbers from the ICD-9 version with the ICD-10 version. The numbers between the 2014 and 2015 linkage cohorts were comparable, so no formal validation was done at that time.

5) 2017 Update-(post EPIC-Implementation):
In January of 2016 the institution adopted the EPIC electronic medical record. Prior to EPIC, problem list terms were generated separately from ICD-9/10 diagnostic billing terms. In EPIC, problem list terms and diagnostic billing codes are generated as part of the same process. This led to the need to change the algorithm to reflect this lack of independence in choosing problem list terms and billing codes as part of a clinical encounter. Procedure code use remains unchanged in the new algorithms. The new inclusion criteria now include two problem list terms from separate outpatient visits or 1 diagnosis billing code plus a Problem list term also from separate outpatient visits. A blinded list of 656 patients was reviewed for formal validation (10 pts per practice with each condition and 7 pts with none of the conditions).