[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 35.175.190.77. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Figure 1.
Initial Study Cohort
Initial Study Cohort

LHIN indicates Local Health Integration Network; OHIP, Ontario Health Insurance Plan.

Figure 2.
Rates of ECG Ordering by Primary Care Practice
Rates of ECG Ordering by Primary Care Practice
Table 1.  
Patient- and Physician-Level Characteristics Based on ECG Claim Status 30 Days After Patient’s Index Annual Health Examination, n = 3 629 859
Patient- and Physician-Level Characteristics Based on ECG Claim Status 30 Days After Patient’s Index Annual Health Examination, n = 3 629 859
Table 2.  
Association of Patient- and Physician-Level Characteristics With Having a Potentially Low-Value ECG Within 30 Days After an Annual Health Examination in 2 873 357 Patientsa
Association of Patient- and Physician-Level Characteristics With Having a Potentially Low-Value ECG Within 30 Days After an Annual Health Examination in 2 873 357 Patientsa
Table 3.  
Patient- and Physician-Level Indicators for a Downstream Cardiac Consultation, Test, or Procedure in 2 352 324 Patients Within 3 Months After an Annual Health Examination Based on a Multilevel Logistic Regression With a Random Intercept for Practice-Level Effects
Patient- and Physician-Level Indicators for a Downstream Cardiac Consultation, Test, or Procedure in 2 352 324 Patients Within 3 Months After an Annual Health Examination Based on a Multilevel Logistic Regression With a Random Intercept for Practice-Level Effects
1.
Institute of Medicine.  Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy Press; 2001.
2.
Wennberg  JE, Fisher  ES, Skinner  JS.  Geography and the debate over Medicare reform.  Health Aff (Millwood). 2002;(suppl web exclusives):W96-114.PubMedGoogle Scholar
3.
Levinson  W, Kallewaard  M, Bhatia  RS, Wolfson  D, Shortt  S, Kerr  EA; Choosing Wisely International Working Group.  ‘Choosing Wisely’: a growing international campaign.  BMJ Qual Saf. 2015;24(2):167-174.PubMedGoogle ScholarCrossref
4.
Chou  R, Arora  B, Dana  T, Fu  R, Walker  M, Humphrey  L.  Screening asymptomatic adults with resting or exercise electrocardiography: a review of the evidence for the U.S. Preventive Services Task Force.  Ann Intern Med. 2011;155(6):375-385.PubMedGoogle ScholarCrossref
5.
Wolfson  D, Santa  J, Slass  L.  Engaging physicians and consumers in conversations about treatment overuse and waste: a short history of the choosing wisely campaign.  Acad Med. 2014;89(7):990-995.PubMedGoogle ScholarCrossref
6.
Bhatia  RS, Levinson  W, Shortt  S,  et al.  Measuring the effect of Choosing Wisely: an integrated framework to assess campaign impact on low-value care.  BMJ Qual Saf. 2015;24(8):523-531.PubMedGoogle ScholarCrossref
7.
Colla  CH, Sequist  TD, Rosenthal  MB, Schpero  WL, Gottlieb  DJ, Morden  NE.  Use of non-indicated cardiac testing in low-risk patients: Choosing Wisely.  BMJ Qual Saf. 2015;24(2):149-153.PubMedGoogle ScholarCrossref
8.
Rosenberg  A, Agiro  A, Gottlieb  M,  et al.  Early trends among seven recommendations from the Choosing Wisely Campaign.  JAMA Intern Med. 2015;175(12):1913-1920.PubMedGoogle ScholarCrossref
9.
Kale  MS, Bishop  TF, Federman  AD, Keyhani  S.  Trends in the overuse of ambulatory health care services in the United States.  JAMA Intern Med. 2013;173(2):142-148.PubMedGoogle ScholarCrossref
10.
Blecker  S, Bhatia  RS, You  JJ,  et al.  Temporal trends in the utilization of echocardiography in Ontario, 2001 to 2009.  JACC Cardiovasc Imaging. 2013;6(4):515-522.PubMedGoogle ScholarCrossref
11.
Lucas  FL, DeLorenzo  MA, Siewers  AE, Wennberg  DE.  Temporal trends in the utilization of diagnostic testing and treatments for cardiovascular disease in the United States, 1993-2001.  Circulation. 2006;113(3):374-379.PubMedGoogle ScholarCrossref
12.
The College of Family Physicians of Canada (CFPC). Annual physical examination practices by province/territory in Canada. CFPC; 2013 Nov. http://www.cfpc.ca/CFPC-PT-Annual-Exam. Accessed January 2, 2017.
13.
Government of Ontario. Periodic personal health visit. Government of Ontario; 2013. http://www.health.gov.on.ca/en/pro/programs/phys_services/docs/periodic_health_visit_is_ea_en.pdf. Accessed January 2, 2017.
14.
Colla  CH, Morden  NE, Sequist  TD, Schpero  WL, Rosenthal  MB.  Choosing wisely: prevalence and correlates of low-value health care services in the United States.  J Gen Intern Med. 2015;30(2):221-228.PubMedGoogle ScholarCrossref
15.
Schwartz  AL, Landon  BE, Elshaug  AG, Chernew  ME, McWilliams  JM.  Measuring low-value care in Medicare.  JAMA Intern Med. 2014;174(7):1067-1076.PubMedGoogle ScholarCrossref
16.
Statistics Canada. Postal Code Conversion File Plus (PCCF+), reference guide. Statistics Canada; 2013. http://data.library.utoronto.ca/datapub/codebooks/cstdli/pccf_health/pccf6a1/82-F0086-XDB-2014v6a-eng.pdf. Accessed December 17, 2016.
17.
The College of Family Physicians (CFPC). Best advice: Patient rostering in family practice. CFPC; Nov 2012. http://www.patientsmedicalhome.ca/files/uploads/PMH_Best_Advice_Rostering.pdf. Accessed January 21, 2017.
18.
Kiran  T, Kopp  A, Moineddin  R, Glazier  RH.  Longitudinal evaluation of physician payment reform and team-based care for chronic disease management and prevention.  CMAJ. 2015;187(17):E494-E502.PubMedGoogle ScholarCrossref
19.
Glazier  RH, Hutchinson  B, Kopp  A. Comparison of family health teams to other primary care models, 2004/05 to 2011/12. Institute for Clinical Evaluative Sciences (ICES); November 2015. https://www.ices.on.ca/Publications/Atlases-and-Reports/2015/Comparison-of-Family-Health-Teams. Accessed April 14, 2017.
20.
Bonett  DG.  Confidence interval for a coefficient of quartile variation.  Comput Stat Data Anal. 2006;50(11):2953-2957.Google ScholarCrossref
21.
Larsen  K, Merlo  J.  Appropriate assessment of neighborhood effects on individual health: integrating random and fixed effects in multilevel logistic regression.  Am J Epidemiol. 2005;161(1):81-88.PubMedGoogle ScholarCrossref
22.
Snijders  T A B, Bosker  RJ.  Multilevel analysis: an introduction to basic and advanced multilevel modeling. Thousand Oaks, CA: Sage; 1999.
23.
Merlo  J, Chaix  B, Ohlsson  H,  et al.  A brief conceptual tutorial of multilevel analysis in social epidemiology: using measures of clustering in multilevel logistic regression to investigate contextual phenomena.  J Epidemiol Community Health. 2006;60(4):290-297.PubMedGoogle ScholarCrossref
24.
Pendrith  C, Bhatia  M, Ivers  NM,  et al.  Frequency of and variation in low-value care in primary care: a retrospective cohort study.  CMAJ Open. 2017;5(1):E45-E51.PubMedGoogle ScholarCrossref
25.
Kirkham  KR, Wijeysundera  DN, Pendrith  C,  et al.  Preoperative testing before low-risk surgical procedures.  CMAJ. 2015;187(11):E349-E358. doi:10.1503/cmaj.150174PubMedGoogle ScholarCrossref
26.
Shah  BR, McCoy  LA, Federspiel  JJ,  et al.  Use of stress testing and diagnostic catheterization after coronary stenting: association of site-level patterns with patient characteristics and outcomes in 247,052 Medicare beneficiaries.  J Am Coll Cardiol. 2013;62(5):439-446.PubMedGoogle ScholarCrossref
27.
Wennberg  DE, Kellett  MA, Dickens  JD, Malenka  DJ, Keilson  LM, Keller  RB.  The association between local diagnostic testing intensity and invasive cardiac procedures.  JAMA. 1996;275(15):1161-1164.PubMedGoogle ScholarCrossref
28.
Lucas  FL, Siewers  AE, Malenka  DJ, Wennberg  DE.  Diagnostic-therapeutic cascade revisited: coronary angiography, coronary artery bypass graft surgery, and percutaneous coronary intervention in the modern era.  Circulation. 2008;118(25):2797-2802.PubMedGoogle ScholarCrossref
29.
Shah  BR, Cowper  PA, O’Brien  SM,  et al.  Patterns of cardiac stress testing after revascularization in community practice.  J Am Coll Cardiol. 2010;56(16):1328-1334.PubMedGoogle ScholarCrossref
30.
Ashley  EA, Raxwal  V, Froelicher  V.  An evidence-based review of the resting electrocardiogram as a screening technique for heart disease.  Prog Cardiovasc Dis. 2001;44(1):55-67.PubMedGoogle ScholarCrossref
31.
United States Preventative Services Task Force (USPSTF). Coronary heart disease: screening with electrocardiography. USPSTF; 2012. https://www.uspreventiveservicestaskforce.org/Page/Document/UpdateSummaryFinal/coronary-heart-disease-screening-with-electrocardiography. Accessed January 4, 2017.
32.
Colla  CH.  Swimming against the current—what might work to reduce low-value care?  N Engl J Med. 2014;371(14):1280-1283.PubMedGoogle ScholarCrossref
33.
Greenland  P, Alpert  JS, Beller  GA,  et al; American College of Cardiology Foundation/American Heart Asscoiation Tasl Force on Practice Guidelines.  2010 ACCF/AHA guideline for assessment of cardiovascular risk in asymptomatic adults: executive summary: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines.  Circulation. 2010;122(25):2748-2764.PubMedGoogle ScholarCrossref
34.
Chaudhuri  D, Montgomery  A, Gulenchyn  K, Mitchell  M, Joseph  P.  Effectiveness of quality improvement interventions at reducing inappropriate cardiac imaging: a systematic review and meta-analysis.  Circ Cardiovasc Qual Outcomes. 2016;9(1):7-13.PubMedGoogle ScholarCrossref
35.
Ivers  N, Jamtvedt  G, Flottorp  S,  et al. Audit and feedback: effects on professional practice and patient outcomes. Cochrane Database Syst Rev. http://www.cochrane.org/CD000259/EPOC_audit-and-feedback-effects-on-professional-practice-and-patient-outcomes. Accessed January 5, 2017.
36.
Dudzinski  DM, Bhatia  RS, Mi  MY, Isselbacher  EM, Picard  MH, Weiner  RB.  Effect of an educational intervention on the rate of rarely appropriate outpatient echocardiograms ordered by attending academic cardiologists: a randomized clinical trial.  JAMA Cardiol. 2016;1(7):805-812.PubMedGoogle ScholarCrossref
37.
Jaglal  SB, McIsaac  WJ, Hawker  G, Jaakkimainen  L, Cadarette  SM, Chan  BT.  Patterns of use of the bone mineral density test in Ontario, 1992-1998.  CMAJ. 2000;163(9):1139-1143.PubMedGoogle Scholar
Original Investigation
Less Is More
September 2017

Electrocardiograms in Low-Risk Patients Undergoing an Annual Health Examination

Author Affiliations
  • 1Institute for Health Systems Solutions and Virtual Care, Women’s College Hospital, Toronto, Ontario, Canada
  • 2Institute for Clinical Evaluative Sciences (ICES), Toronto, Ontario, Canada
  • 3Institute for Health Policy, Management, and Evaluation (IHPME), University of Toronto, Toronto, Ontario, Canada
  • 4Department of Family and Community Medicine, University of Toronto, Toronto, Ontario, Canada
  • 5University of Ottawa Medical School, Ottawa, Ontario, Canada
  • 6Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada
  • 7Schulich Heart Center, Sunnybrook Health Sciences Center, University of Toronto, Toronto, Ontario, Canada
  • 8Department of Medicine, University of Toronto, Toronto, Ontario, Canada
  • 9Sunnybrook Research Institute (SRI), Sunnybrook Health Sciences Center, Toronto, Ontario, Canada
  • 10Department of Family and Community Medicine, Women’s College Hospital, Toronto, Ontario, Canada
  • 11Division of Cardiology, University of Toronto, Toronto, Ontario, Canada
  • 12Keenan Research Centre for Biomedical Science, St Michael’s Hospital, Toronto, Ontario, Canada
  • 13Li Ka Shing Knowledge Institute, St Michael’s Hospital, Toronto, Ontario, Canada
  • 14Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, St Michael’s Hospital, Toronto, Ontario, Canada
  • 15Department of Family and Community Medicine, St Michael’s Hospital, Toronto, Ontario, Canada
  • 16Department of Medicine, University of Toronto, Toronto, Ontario, Canada
JAMA Intern Med. 2017;177(9):1326-1333. doi:10.1001/jamainternmed.2017.2649
Key Points

Question  How frequently are electrocardiograms (ECG) ordered as part of an annual health examination (AHE) and what is the impact of an ECG on downstream cardiac testing?

Findings  In this population-based cohort study of 3 629 859 adult patients who had an AHE, 21.5% had an ECG within 30 days of an AHE. Those who had an ECG were 5 times more likely to have another cardiac test or consultation than those who did not.

Meaning  Electrocardiograms conducted on patients as part of an AHE are common and are associated with more cardiac tests and consultations.

Abstract

Importance  Clinical guidelines advise against routine electrocardiograms (ECG) in low-risk, asymptomatic patients, but the frequency and impact of such ECGs are unknown.

Objective  To assess the frequency of ECGs following an annual health examination (AHE) with a primary care physician among patients with no known cardiac conditions or risk factors, to explore factors predictive of receiving an ECG in this clinical scenario, and to compare downstream cardiac testing and clinical outcomes in low-risk patients who did and did not receive an ECG after their AHE.

Design, Setting, and Participants  A population-based retrospective cohort study using administrative health care databases from Ontario, Canada, between 2010/2011 and 2014/2015 to identify low-risk primary care patients and to assess the subsequent outcomes of interest in this time frame. All patients 18 years or older who had no prior cardiac medical history or risk factors who received an AHE.

Exposures  Receipt of an ECG within 30 days of an AHE.

Main Outcomes and Measures  Primary outcome was receipt of downstream cardiac testing or consultation with a cardiologist. Secondary outcomes were death, hospitalization, and revascularization at 12 months.

Results  A total of 3 629 859 adult patients had at least 1 AHE between fiscal years 2010/2011 and 2014/2015. Of these patients, 21.5% had an ECG within 30 days after an AHE. The proportion of patients receiving an ECG after an AHE varied from 1.8% to 76.1% among 679 primary care practices (coefficient of quartile dispersion [CQD], 0.50) and from 1.1% to 94.9% among 8036 primary care physicians (CQD, 0.54). Patients who had an ECG were significantly more likely to receive additional cardiac tests, visits, or procedures than those who did not (odds ratio [OR], 5.14; 95% CI, 5.07-5.21; P < .001). The rates of death (0.19% vs 0.16%), cardiac-related hospitalizations (0.46% vs 0.12%), and coronary revascularizations (0.20% vs 0.04%) were low in both the ECG and non-ECG cohorts.

Conclusions and Relevance  Despite recommendations to the contrary, ECG testing after an AHE is relatively common, with significant variation among primary care physicians. Routine ECG testing seems to increase risk for a subsequent cardiology testing and consultation cascade, even though the overall cardiac event rate in both groups was very low.

Introduction

Low-value care, defined as care where there is a lack of benefit or where the benefits are outweighed by the potential risks, can lead to higher health care costs, patient inconvenience, and in some cases harm to patients.1-3 Resting electrocardiography (ECG) in low-risk patients undergoing an annual health examination (AHE) by a primary care physician is an example of low-value care. In 2012, the United States Preventive Services Task Force (USPSTF) recommended against routine ECG screening in low-risk patients because there is inadequate evidence for the added utility of ECG in the diagnosis of coronary disease.4 The Choosing Wisely campaign, launched in 2012, also saw multiple specialty societies include a recommendation against noninvasive cardiac testing in low-risk or asymptomatic patients in their top 5 lists of low-value tests, treatments, and procedures that physicians and patients should question.5

As interest in curbing low-value care increases, accurate estimates of the utilization of low-value care and its costs and impact on patient outcomes are vital.6 While prior research has estimated the frequency of ECGs in selected low-risk patient groups, population-wide usage of ECGs in low-risk patients and in particular its impact on costs and cardiovascular outcomes is unknown.7-9 Understanding the association between low-value cardiac testing and subsequent health care utilization and outcomes is essential in the face of concerns regarding rising cardiovascular testing utilization.10,11

The aim of this study is to quantify the frequency of ECGs ordered after an AHE in low-risk primary care patients with no prior cardiac medical history. In addition, we examined whether such ECGs are associated with subsequent cardiac tests or consultations and/or patient outcomes.

Methods
Study Design and Data Sources

We conducted a retrospective cohort study in Ontario, Canada, using linked population-based administrative health care databases. The data sets were linked using unique encoded identifiers and analyzed at the Institute for Clinical Evaluative Sciences (ICES). Patients were included if they were an Ontario resident with a valid health card number and had at least 1 claim for an annual health examination (AHE) with a primary care physician (ie, a family physician), between April 1, 2010, and March 31, 2015. Index AHE claims were identified via Ontario Health Insurance Plan (OHIP) billing codes for either a periodic health visit (PHV) or an AHE conducted on a healthy adult patient (aged 18 years or older).12 The general AHE code was replaced by the PHV in 2013.13 Similar to an AHE with a healthy patient, the newer PHV is defined as a service performed on “healthy patients who have no apparent medical problems. The physician and patient can use the appointment to discuss prevention like screening for cancer and other health issues relevant to the individual patient’s medical history and lifestyle.”13 Beyond the PHV having differential billing codes for different aged patients, there are no practical differences between the AHE and PHV,13 and so both were considered interchangeable for this anaylsis. We excluded patients residing in long-term care or with incomplete demographic data. Patients with physician visits, emergency department visits, or hospitalizations suggesting significant cardiovascular disease or cardiovascular risk within 3 years prior to their AHE were deemed high-risk (eg, prior myocardial infarction, hypertension, diabetes) and excluded through application of diagnostic codes from OHIP and International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes from the Discharge Abstract Database (DAD).7,14,15 Patients with prior cardiac-related procedures (eg, aortic valve replacement, coronary artery revascularization), hospitalizations, or consultations were also excluded using ICD-10 and Canadian Classification of Health Interventions (CCI) codes.14 We excluded patients with a history of diabetes and hypertension using the Ontario Diabetes Database and the Ontario Hypertension Database. eAppendix 1 in the Supplement details the full extent of inclusion and exclusion criteria, including data sources and associated codes.

Several data sources were used to define patient- and physician-level characteristics to describe the patient cohort and create multivariable statistical models. Demographic data on patients (ie, age, sex, rurality), were obtained from the Registered Persons Database (RDPB). Quintiles of median neighborhood income were used to approximate patients’ socioeconomic status.16 Any hospitalizations (except those for cardiac-related reasons) within 3 years prior of the index examination were determined from DAD ICD-10 codes. OHIP and DAD were used to determine patients’ history of cancer, chronic obstructive pulmonary disease (COPD), asthma, dementia, mental illness, and rheumatoid arthritis within 3 years prior to study entry. Whether or not a patient was rostered to a regular primary care physician was determined through cross-referencing Client Agency Program Enrolment (CAPE) tables and OHIP fee codes. Ontario introduced formal rostering in its primary care patient enrolment models in 2001 and during the study period most (70%) Ontarian patients were rostered.17 Physician-level variables were primarily identified by linking physicians from patients’ OHIP claims with the ICES Physicians Database (IPDB) to determine physician sex, years since graduation, and international medical graduate status. The CAPE tables and OHIP claims were used to classify primary care practice groups (hereby referred to as practices).18,19 These practices consisted of 3 or more physicians submitting joint claims to the Ministry of Health and Long Term Care for reimbursement (ie, billing groups).19 Payment models were noted and any practices consisting of fewer than 3 physicians were excluded for privacy reasons.

Research ethics approval was received from Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada. Patient consent was waived because all data were deidentified.

Low-Value Cardiac Care

The primary outcome was low-value cardiac care, defined as patient receipt of at least 1 ECG within 30 days of a patient’s index AHE.8 Ontario Health Insurance Plan claims were used to identify any electrocardiography performed on the patient in that timeframe.

Downstream cardiac care was identified using OHIP claims. Receipt of either a consultation with a cardiologist or cardiac surgeon, transthoracic echocardiogram (TTE), stress test, nuclear stress test, and/or cardiac catheterization procedure were measured where claims occurred within 90 days of an eligible AHE.

Cardiac outcomes, including death, cardiac-related hospitalizations, and coronary revascularization procedures were captured within 12 months of the index AHE. Death was identified via the RPBD, whereas the latter events were identified using DAD ICD-10 codes and CCI codes, respectively. Percutaneous coronary interventions (PCI) and coronary artery bypass grafts (CABG) were also identified using OHIP claims data.

Statistical Analysis

Unadjusted 30-day ECG rates were calculated separately at the level of individual regions, practices, and physicians between fiscal years 2010/2011 and 2014/2015. Regions were defined by Ontario’s 14 Local Health Integration Networks (LHINs)—geographically organized administrative regions that plan, integrate, and fund local health care. Variation in ordering rates was assessed at each level by calculating the coefficient of quartile deviation, a potentially more robust alternative to the coefficient of variation.20 All statistical analyses were performed using SAS statistical software (version 9.4, SAS Institute).

Hierarchical random-intercept multivariable logistic regression models were used to assess the association of patient- and physician-level characteristics with the occurrence of an ECG within 30 days post-AHE. Owing to the physician-level characteristics included in these models, only patients who could be linked to a family physician belonging to an identifiable practice were included. To test the sensitivity of our primary outcome definition, we ran a similar analysis including only patients that had an ECG on the same day as an AHE.

Through inclusion of random intercepts for each practice, it was possible to calculate the median odds ratio (MOR), a measure of the heterogeneity in ordering a post-AHE ECG among practices.21-23 If one were to repeatedly sample 2 participants with the same covariates from different practices randomly, then the MOR is the relative odds difference between the participant at higher risk of receiving an ECG and the participant at the lower risk of receiving an ECG in the median case.21-23 The MOR always has a value greater than or equal to 1 because it places the patient belonging to the practice with higher odds of ordering an ECG in the numerator.21-23 The value is directly comparable to a fixed effects OR, and is based on a between-practice variance estimate that is adjusted for all other factors present in a multilevel model.21 For example, a MOR of 2.00 suggests, in the median case, 100% higher odds of receiving an ECG at 1 practice vs another.23 The intracluster correlation coefficient (ICC) was also calculated using the linear threshold method to estimate the proportion of total variance in ECG ordering that can be attributed to between-practice differences.21-23

In addition, hierarchical random-intercept multivariable logistic regression models were used to assess the impact of having postvisit ECG on downstream cardiac care and adverse patient outcomes. For each of these additional models, receipt of an ECG within 30 days post-AHE was the primary exposure of interest and included as a covariate. To account for the possibility that a physician’s propensity to order low-value ECGs may be associated with the propensity to order downstream care we determined each physician’s ECG ordering rate (ie, the proportion of patient AHEs with an ECG ordered within 30 days) and then created a variable that stratified physicians into quintiles based on their individual rates. This factor representing physician ECG ordering quintile was then included in the regression models for both downstream care and outcomes.

Results
ECG Status and Corresponding Characteristics

The study cohort consisted of 3 629 859 adult patients with at least 1 AHE between fiscal years 2010/2011 and 2014/2015 in Ontario (Figure 1). Demographic and clinical data for all eligible patients are presented in Table 1, along with characteristics of their corresponding primary care physicians and practice groups. Overall, 21.5% of adult patients had a potentially low-value ECG within 30 days of their index AHE, 51.7% of ECG claims occurred on the same day as the AHE, and 78.5% of ECG claims were within 7 days of an AHE. Patients who had at least 1 ECG in the defined follow-up period were more likely to be male and were generally older than those who did not have an ECG. The comorbidity burden was low for the entire cohort, though there were some statistically significant differences between groups.

Variation by Region, Practice, and Physician

Regional variation of ECG ordering ranged from a low of 0.7% in the North West LHIN to 24.4% in the Central LHIN (coefficient of quartile deviation [CQD], 0.71) (eAppendix 2.0 in the Supplement). Among 679 practices, the proportion of patients who received an ECG post-AHE ranged from 1.8% to 76.1% (CQD, 0.50) as shown in Figure 2. Substantial variation was also observed across individual primary care physicians (range, 1.1%-94.9%; CQD, 0.54) as shown in eAppendix 2.1 in the Supplement. Among the 8036 primary care physicians included, 7.2% ordered ECGs on more than 50% of their patients following an AHE.

Factors Associated With Ordering an ECG Post-AHE

From the initial study cohort detailed in Figure 1, a total of 2 873 357 adult patients who reported belonging to, and could be linked to, an established practice were eligible for modeling of post-AHE ECG receipt status.

As described in Table 2, older age was associated with increased odds of having an ECG within 30 days after an AHE. Patients living in a rural area were less likely to have an ECG than those living in urban areas. Patients with rheumatological disease and cancer had increased odds of having an ECG. Physician factors associated with ECG ordering were male sex, international medical graduate status, and having practiced for longer than 30 years. The interpractice variation in ECG ordering was significant. The MOR was 2.50, indicating that the odds of a patient having a post-AHE ECG at 1 randomly selected high-ordering practice were 150% greater than a patient with the same characteristics at another randomly selected, low-ordering practice. Aside from patient age, the effect of a patients’ practice membership on receipt of an ECG was stronger than the association observed between the outcome and any other patient- or physician-level factors. The ICC estimate indicates that 21.9% of the total variation in post-AHE ECG use can be attributed to practice-level variation. Based on the sensitivity analysis, all of the characteristics significantly associated with increased odds of having an ECG at 30 days post-AHE were similarly associated with same-day receipt of an ECG (eAppendix 2.2 in the Supplement).

Association of Post-AHE ECG With Downstream Cardiac Care

Overall, 5.3% of patients with an AHE had a cardiac-related consultation, test, or procedure within 90 days following their primary care physician AHE. Cardiac care included consultations with a cardiologist or cardiac surgeon (n = 36 085), transthoracic echocardiograms (TTE, n = 83 463), stress tests (n = 42 923), nuclear stress tests (n = 15 651), and cardiac catheterizations (n = 1830).

Table 3 presents the final multivariable models for downstream cardiac care. After adjusting for physician ECG ordering quintile, patients with an ECG within 30 days of an AHE had significantly higher odds of having further cardiac tests or consultations compared with those who did not. There was significant interpractice variation in ordering downstream care (cardiac consultations: MOR, 1.47; ICC, 4.7%; TTEs: MOR, 1.60; ICC, 6.85%), and either form of stress test (stress test: MOR, 1.71; ICC, 8.85%; nuclear stress test: MOR, 1.56; ICC, 6.23%). Cardiac catheterization rate was low (0.29% vs non-ECG 0.03%) and attempts to model resulted in nonconvergence in the statistical software (eAppendix 2.3 in the Supplement). Sensitivity analyses for ECGs conducted on the same day showed a similar association with higher odds of ordering additional cardiac testing within 90 days (eAppendix 2.4 in the Supplement).

The overall rates of adverse clinical outcomes at 1-year post-AHE were extremely low in both groups. The unadjusted rate of each outcome was higher in the ECG vs the non-ECG group, including death (0.19% vs 0.16%), cardiac-related hospitalizations (0.46% vs 0.12%), and coronary revascularizations (0.20% vs 0.04%) (eAppendix 2.5 in the Supplement). Only the descriptive statistics are presented for these outcomes because their corresponding regression models each failed to converge in the statistical software.

Discussion

In this large, retrospective cohort study, we found that despite recommendations against ECGs in low-risk patients undergoing a routine AHE with a primary care physician, this practice seemed to be common. Importantly, we demonstrated that patients who received an ECG 30 days post-AHE were more than 5 times more likely to also receive another cardiac test, procedure, or consultation with a specialist. These downstream cardiac tests and procedures also demonstrated significant practice variation even after adjustment for patient and physician factors. Finally, both groups of patients exhibited low combined rates of death, cardiac-related hospitalizations, and coronary revascularizations in the ensuing year.

These findings support earlier results that show high rates of potentially low-value ECGs in primary care patients. Previous studies have shown the frequency of ECGs in low-risk patients to be between 9% to 12% in various populations, including both Medicare and commercial payers, with similar degrees of regional variation to this study.7-9 Where this study differs is that it includes an entire population from a single-payer, publicly funded system, allowing for a more robust analysis of ordering practices. We also found significant variation in ordering between regions, practices, and even physicians that cannot be explained by patient factors such as comorbidities. This striking ordering variation, which has been noted in both primary care and hospital-based ordering practices in previous work, provides potential opportunities for improvements in ordering practices, particularly among high-ordering physicians, because there were a small number of physicians that ordered ECGs on most of their patients.24,25 Most importantly, we demonstrate that ECGs in this low-risk population leads to further downstream cardiac testing and consultations that add to health care costs.

One of the most important findings of this study was the higher rates of further cardiac testing or cardiology consultations in patients who had an ECG. The diagnostic cascade is a described phenomenon where higher rates of noninvasive diagnostic testing lead to higher rates of more invasive diagnostic testing and therapeutic interventions.26-29 For example, Shah and colleagues26 found higher rates of routine stress testing after coronary revascularization lead to higher rates of repeated revascularization with no impact on death or repeated myocardial infarction. To date, however, this finding has not been shown in resting ECG testing, which is most commonly done in primary care offices in patients who often do not have a medical history of cardiac disease. The higher rates of further testing or consultations, like ECG ordering, were independent of patient comorbidities; it is probable that incidentally discovered abnormalities found on ECG precipitated further testing and cardiology consultations.

The low event rates in patients who received and did not receive an ECG as part of an AHE add more evidence to the recommendation against using ECGs as a risk stratification tool. Systematic reviews have found no randomized clinical trials or large prospective cohort studies on the effects of ECGs ordered in low-risk patients vs no ECG on clinical outcomes or costs.4,30 Despite this lack of evidence there still remains some debate as to the utility of ECGs as a risk stratification tool.29 Our findings in a large, population-based cohort study show very low event rates that were less than 1% in both groups, despite higher rates of cardiovascular investigations in the ECG group. This data lends further evidence to the current guidelines recommending against routine ECGs in low-risk patients, which appear to lead to higher health care utilization with questionable clinical benefit.

The results of this study have considerable health care policy implications. First, when selecting overuse metrics for quality improvement initiatives, consideration should be given to the impact on downstream testing and outcomes.31,32 Some of the past criticism of the Choosing Wisely recommendations is that they are often of seemingly little consequence, with specialty societies avoiding big-ticket items like surgical procedures, or more advanced diagnostic tests.33 Our findings suggest that even low-cost procedures, like ECGs in low-risk patients occur with considerable frequency, and importantly can lead to more advanced testing that adds costs with little potential benefit to patients. Second, measurement of low-value care should also attempt to quantify the impact on health outcomes for patients.6 Finally, quality improvement interventions to reduce low-value care could be designed to more effectively target practices and physicians with high ordering rates to reduce the prevalence of low-value cardiac testing, and the unexplained ordering variation.34 In particular, prior research has shown that the use of audit and feedback, decision support tools, and education could reduce low-value cardiac testing, and similar types of interventions could be used in some combination to reduce low-value care.24,35-37

Limitations

The results must be interpreted in the context of several limitations. Administrative databases do not provide important clinical information, such as presence of symptoms or abnormal physical examination findings that are necessary to determine appropriateness of further testing orders. In this instance, we are unable to determine whether a patient had an ECG for screening, or if the patient had symptoms or signs that warranted a diagnostic test. However, usually a visit will not be coded as an AHE if there is a specific cardiac symptom that is the focus of the assessment. It is also possible that some physicians ordered an ECG prior to the AHE that were not captured. We also do not have the results of the ECGs, which would almost certainly have influenced further cardiac investigations. Finally, it is possible that practice or regional factors not identified by available data may play a role in the degree of ordering variation seen, which may be an opportunity for future studies. Despite these limitations, this study provides important new information about the use and impact of low-value cardiac testing in the primary care setting.

Conclusions

In this large, population-based retrospective study, we found 21.5% of low-risk patients received an ECG within 30 days following an AHE, with significant regional-, practice-, and physician-level ordering variation. Moreover, low-risk patients who received an ECG also had a higher likelihood of further cardiac tests, procedures, and cardiologist consultations.

Back to top
Article Information

Corresponding Author: R. Sacha Bhatia, MD, MBA, Institute for Health Systems Solutions and Virtual Care, Women’s College Hospital, 76 Grenville St, 6th Floor, Toronto, ON M5S 1B1, Canada (sacha.bhatia@wchospital.ca).

Accepted for Publication: May 1, 2017.

Published Online: July 10, 2017. doi:10.1001/jamainternmed.2017.2649

Author Contributions: Dr Bhatia had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Bhatia, Bouck, Pendrith, Ivers, Tu, Wintemute, Dorian, Tepper, Glazier, Levinson.

Acquisition, analysis, or interpretation of data: Bhatia, Bouck, Ivers, Singh, Pendrith, Mecredy, Ko, Martin, Wilson, Dorian, Tepper, Austin, Glazier.

Drafting of the manuscript: Bhatia, Bouck, Ivers, Singh, Pendrith, Wilson.

Critical revision of the manuscript for important intellectual content: Bhatia, Bouck, Ivers, Pendrith, Mecredy, Ko, Martin, Wijeysundera, Tu, Wintemute, Dorian, Tepper, Austin, Glazier, Levinson.

Statistical analysis: Bhatia, Bouck, Ivers, Singh, Mecredy, Wijeysundera.

Obtained funding: Bhatia, Levinson.

Administrative, technical, or material support: Bhatia, Ivers, Pendrith, Tu, Wilson, Dorian.

Study supervision: Bhatia, Dorian, Levinson.

Conflict of Interest Disclosures: Dr Ivers is supported by New Investigator Awards from the Canadian Institutes of Health Research and the Department of Family and Community Medicine, University of Toronto. Dr Wijeysundera is supported by the Distinguished Clinician Scientist Award from the Heart and Stroke Foundation of Canada. Dr Glazier is supported as a Clinician scientist in the Department of family and Community Medicine at St Michael’s Hospital and at the University of Toronto. Dr Ko is supported by a Mid-Career Investigator Award from the Heart and Stroke Foundation, Ontario Provincial Office. Jasjit Singh was supported by a CIHR sponsored CANHEART summer student grant. Dr Austin is supported by a Career Investigator Award from the Heart and Stroke Foundation, Ontario Provincial Office. No other disclosures are reported.

Funding/Support: This study was supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded in part by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; the collection, management, analysis, or interpretation of the data; the preparation, review or approval of the manuscript; or the decision to submit the manuscript for publication.

Disclaimer: The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred. Parts of this material are based on data and information compiled and provided by CIHI. However, the analyses, conclusions, opinions, and statements expressed herein are those of the authors, and not necessarily those of CIHI.

References
1.
Institute of Medicine.  Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy Press; 2001.
2.
Wennberg  JE, Fisher  ES, Skinner  JS.  Geography and the debate over Medicare reform.  Health Aff (Millwood). 2002;(suppl web exclusives):W96-114.PubMedGoogle Scholar
3.
Levinson  W, Kallewaard  M, Bhatia  RS, Wolfson  D, Shortt  S, Kerr  EA; Choosing Wisely International Working Group.  ‘Choosing Wisely’: a growing international campaign.  BMJ Qual Saf. 2015;24(2):167-174.PubMedGoogle ScholarCrossref
4.
Chou  R, Arora  B, Dana  T, Fu  R, Walker  M, Humphrey  L.  Screening asymptomatic adults with resting or exercise electrocardiography: a review of the evidence for the U.S. Preventive Services Task Force.  Ann Intern Med. 2011;155(6):375-385.PubMedGoogle ScholarCrossref
5.
Wolfson  D, Santa  J, Slass  L.  Engaging physicians and consumers in conversations about treatment overuse and waste: a short history of the choosing wisely campaign.  Acad Med. 2014;89(7):990-995.PubMedGoogle ScholarCrossref
6.
Bhatia  RS, Levinson  W, Shortt  S,  et al.  Measuring the effect of Choosing Wisely: an integrated framework to assess campaign impact on low-value care.  BMJ Qual Saf. 2015;24(8):523-531.PubMedGoogle ScholarCrossref
7.
Colla  CH, Sequist  TD, Rosenthal  MB, Schpero  WL, Gottlieb  DJ, Morden  NE.  Use of non-indicated cardiac testing in low-risk patients: Choosing Wisely.  BMJ Qual Saf. 2015;24(2):149-153.PubMedGoogle ScholarCrossref
8.
Rosenberg  A, Agiro  A, Gottlieb  M,  et al.  Early trends among seven recommendations from the Choosing Wisely Campaign.  JAMA Intern Med. 2015;175(12):1913-1920.PubMedGoogle ScholarCrossref
9.
Kale  MS, Bishop  TF, Federman  AD, Keyhani  S.  Trends in the overuse of ambulatory health care services in the United States.  JAMA Intern Med. 2013;173(2):142-148.PubMedGoogle ScholarCrossref
10.
Blecker  S, Bhatia  RS, You  JJ,  et al.  Temporal trends in the utilization of echocardiography in Ontario, 2001 to 2009.  JACC Cardiovasc Imaging. 2013;6(4):515-522.PubMedGoogle ScholarCrossref
11.
Lucas  FL, DeLorenzo  MA, Siewers  AE, Wennberg  DE.  Temporal trends in the utilization of diagnostic testing and treatments for cardiovascular disease in the United States, 1993-2001.  Circulation. 2006;113(3):374-379.PubMedGoogle ScholarCrossref
12.
The College of Family Physicians of Canada (CFPC). Annual physical examination practices by province/territory in Canada. CFPC; 2013 Nov. http://www.cfpc.ca/CFPC-PT-Annual-Exam. Accessed January 2, 2017.
13.
Government of Ontario. Periodic personal health visit. Government of Ontario; 2013. http://www.health.gov.on.ca/en/pro/programs/phys_services/docs/periodic_health_visit_is_ea_en.pdf. Accessed January 2, 2017.
14.
Colla  CH, Morden  NE, Sequist  TD, Schpero  WL, Rosenthal  MB.  Choosing wisely: prevalence and correlates of low-value health care services in the United States.  J Gen Intern Med. 2015;30(2):221-228.PubMedGoogle ScholarCrossref
15.
Schwartz  AL, Landon  BE, Elshaug  AG, Chernew  ME, McWilliams  JM.  Measuring low-value care in Medicare.  JAMA Intern Med. 2014;174(7):1067-1076.PubMedGoogle ScholarCrossref
16.
Statistics Canada. Postal Code Conversion File Plus (PCCF+), reference guide. Statistics Canada; 2013. http://data.library.utoronto.ca/datapub/codebooks/cstdli/pccf_health/pccf6a1/82-F0086-XDB-2014v6a-eng.pdf. Accessed December 17, 2016.
17.
The College of Family Physicians (CFPC). Best advice: Patient rostering in family practice. CFPC; Nov 2012. http://www.patientsmedicalhome.ca/files/uploads/PMH_Best_Advice_Rostering.pdf. Accessed January 21, 2017.
18.
Kiran  T, Kopp  A, Moineddin  R, Glazier  RH.  Longitudinal evaluation of physician payment reform and team-based care for chronic disease management and prevention.  CMAJ. 2015;187(17):E494-E502.PubMedGoogle ScholarCrossref
19.
Glazier  RH, Hutchinson  B, Kopp  A. Comparison of family health teams to other primary care models, 2004/05 to 2011/12. Institute for Clinical Evaluative Sciences (ICES); November 2015. https://www.ices.on.ca/Publications/Atlases-and-Reports/2015/Comparison-of-Family-Health-Teams. Accessed April 14, 2017.
20.
Bonett  DG.  Confidence interval for a coefficient of quartile variation.  Comput Stat Data Anal. 2006;50(11):2953-2957.Google ScholarCrossref
21.
Larsen  K, Merlo  J.  Appropriate assessment of neighborhood effects on individual health: integrating random and fixed effects in multilevel logistic regression.  Am J Epidemiol. 2005;161(1):81-88.PubMedGoogle ScholarCrossref
22.
Snijders  T A B, Bosker  RJ.  Multilevel analysis: an introduction to basic and advanced multilevel modeling. Thousand Oaks, CA: Sage; 1999.
23.
Merlo  J, Chaix  B, Ohlsson  H,  et al.  A brief conceptual tutorial of multilevel analysis in social epidemiology: using measures of clustering in multilevel logistic regression to investigate contextual phenomena.  J Epidemiol Community Health. 2006;60(4):290-297.PubMedGoogle ScholarCrossref
24.
Pendrith  C, Bhatia  M, Ivers  NM,  et al.  Frequency of and variation in low-value care in primary care: a retrospective cohort study.  CMAJ Open. 2017;5(1):E45-E51.PubMedGoogle ScholarCrossref
25.
Kirkham  KR, Wijeysundera  DN, Pendrith  C,  et al.  Preoperative testing before low-risk surgical procedures.  CMAJ. 2015;187(11):E349-E358. doi:10.1503/cmaj.150174PubMedGoogle ScholarCrossref
26.
Shah  BR, McCoy  LA, Federspiel  JJ,  et al.  Use of stress testing and diagnostic catheterization after coronary stenting: association of site-level patterns with patient characteristics and outcomes in 247,052 Medicare beneficiaries.  J Am Coll Cardiol. 2013;62(5):439-446.PubMedGoogle ScholarCrossref
27.
Wennberg  DE, Kellett  MA, Dickens  JD, Malenka  DJ, Keilson  LM, Keller  RB.  The association between local diagnostic testing intensity and invasive cardiac procedures.  JAMA. 1996;275(15):1161-1164.PubMedGoogle ScholarCrossref
28.
Lucas  FL, Siewers  AE, Malenka  DJ, Wennberg  DE.  Diagnostic-therapeutic cascade revisited: coronary angiography, coronary artery bypass graft surgery, and percutaneous coronary intervention in the modern era.  Circulation. 2008;118(25):2797-2802.PubMedGoogle ScholarCrossref
29.
Shah  BR, Cowper  PA, O’Brien  SM,  et al.  Patterns of cardiac stress testing after revascularization in community practice.  J Am Coll Cardiol. 2010;56(16):1328-1334.PubMedGoogle ScholarCrossref
30.
Ashley  EA, Raxwal  V, Froelicher  V.  An evidence-based review of the resting electrocardiogram as a screening technique for heart disease.  Prog Cardiovasc Dis. 2001;44(1):55-67.PubMedGoogle ScholarCrossref
31.
United States Preventative Services Task Force (USPSTF). Coronary heart disease: screening with electrocardiography. USPSTF; 2012. https://www.uspreventiveservicestaskforce.org/Page/Document/UpdateSummaryFinal/coronary-heart-disease-screening-with-electrocardiography. Accessed January 4, 2017.
32.
Colla  CH.  Swimming against the current—what might work to reduce low-value care?  N Engl J Med. 2014;371(14):1280-1283.PubMedGoogle ScholarCrossref
33.
Greenland  P, Alpert  JS, Beller  GA,  et al; American College of Cardiology Foundation/American Heart Asscoiation Tasl Force on Practice Guidelines.  2010 ACCF/AHA guideline for assessment of cardiovascular risk in asymptomatic adults: executive summary: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines.  Circulation. 2010;122(25):2748-2764.PubMedGoogle ScholarCrossref
34.
Chaudhuri  D, Montgomery  A, Gulenchyn  K, Mitchell  M, Joseph  P.  Effectiveness of quality improvement interventions at reducing inappropriate cardiac imaging: a systematic review and meta-analysis.  Circ Cardiovasc Qual Outcomes. 2016;9(1):7-13.PubMedGoogle ScholarCrossref
35.
Ivers  N, Jamtvedt  G, Flottorp  S,  et al. Audit and feedback: effects on professional practice and patient outcomes. Cochrane Database Syst Rev. http://www.cochrane.org/CD000259/EPOC_audit-and-feedback-effects-on-professional-practice-and-patient-outcomes. Accessed January 5, 2017.
36.
Dudzinski  DM, Bhatia  RS, Mi  MY, Isselbacher  EM, Picard  MH, Weiner  RB.  Effect of an educational intervention on the rate of rarely appropriate outpatient echocardiograms ordered by attending academic cardiologists: a randomized clinical trial.  JAMA Cardiol. 2016;1(7):805-812.PubMedGoogle ScholarCrossref
37.
Jaglal  SB, McIsaac  WJ, Hawker  G, Jaakkimainen  L, Cadarette  SM, Chan  BT.  Patterns of use of the bone mineral density test in Ontario, 1992-1998.  CMAJ. 2000;163(9):1139-1143.PubMedGoogle Scholar
×