Population Trends From 2000-2011 in Nuclear Myocardial Perfusion Imaging Use

Nuclear myocardial perfusion imaging (MPI) accounted for much of the rapid growth in cardiac imaging that occurred from the 1990s through the middle 2000s.^1,2 Factors potentially discouraging use (including publication of appropriate use criteria) have since emerged,³ and recent data reveal modest declines in MPI use in the Medicare fee-for-service population.⁴

We investigated temporal trends in MPI use within a large, community-based population that included persons younger than 65 years and explored whether increasing use of other noninvasive imaging modalities potentially offset declining MPI use.

Patient-level data for MPI performed from 2000-2011 were obtained for members aged 30 years or older from the clinical databases of Kaiser Permanente Northern California, an integrated health care delivery system providing comprehensive inpatient and outpatient care for more than 2.3 million adults. We calculated age- and sex-adjusted annual rates of MPI tests/100 000 person-years using direct adjustment methods with 2011 as the reference year.

Denominators were member-months of membership in individuals with at least 1 month of membership during the year of interest. Linear trends in rates were assessed using the Cochran-Armitage test. Logistic regressions were used to assess interactions of age, sex, setting (outpatient vs inpatient), and prior coronary revascularization (for outpatients) on trends with a random-effects term included to account for clustering by facility. As a surrogate for incident coronary disease, trends in myocardial infarction (MI) were determined using previously described methods.⁵ Two-sided P values of less than .05 were considered significant.

To assess potential substitution of other imaging modalities for MPI, annual rates of cardiac computed tomography and stress echocardiography from 2007-2011 were estimated using a referrals database. Manual audits of randomly selected referrals were performed to determine the positive predictive value (with 95% confidence intervals) of a referral resulting in a test being performed. Use of perfusion positron emission tomography and perfusion magnetic resonance imaging was negligible.

Analyses were performed using SAS version 9.3 (SAS Institute Inc). The institutional review board of the Kaiser Foundation Research Institute approved this study with a waiver of informed consent.

Overall, MPI was used in 302 506 members during 23.2 million person-years of follow-up at 19 facilities. From 2000 until 2006, MPI use increased by 41% (95% CI, 39%-44%; P < .001 for trend) (Figure). In 2006, a reduction began that continued through 2011, with MPI use declining by 51% (95% CI, 50%-52%; P < .001 for trend) (adjusted odds ratio, 0.51 [95% CI, 0.44-0.60] for undergoing MPI in 2011 vs 2006). Relative declines from 2006 to 2011 did not differ by sex or revascularization history, but were greater for outpatients than inpatients (58% vs 31%; P < .001 for interaction) and for persons younger than 65 years vs those aged 65 years or older (56% vs 47%; P < .001 for interaction) (Table).

Stress echocardiography use (tests/100 000 person-years) was unchanged with 189 (95% CI, 180-199) in 2007 and 182 (95% CI, 173-191) in 2011 (P = .93). Cardiac computed tomography use (tests/100 000 person-years) increased from 37 (95% CI, 35-39) in 2007 to 73 (95% CI, 69-77) in 2011 (P = .01), and could have accounted for 5% of the observed decline in overall MPI use if performed as a substitute. During the period of declining MPI use, incident MI declined by 27% (95% CI, 24%-30%; P < .001) in the population from 286 (95% CI, 279-293) events/100 000 person-years to 208 (95% CI, 202-214) events/100 000 person-years.

After increasing from 2000 to 2006, MPI use abruptly declined through 2011 within our population. Declines for persons aged 65 years or older exceeded those for the Medicare fee-for-service population during the same period⁴ and were even greater for younger persons. These declines could not be explained by increasing use of alternative modalities.

Although the abrupt nature of the decline suggests changing physician behavior played a major role, incident coronary disease, as assessed by MI, also declined. We could not determine the relative effects of these factors on MPI use.

Our findings should be interpreted in the context of other limitations. The observed decline occurred in the context of a health care delivery system without direct financial incentives to perform tests. Nevertheless, the substantial reduction in MPI use demonstrates the ability to reduce testing on a large scale with anticipated reductions in health care costs.

Corresponding Author: Edward J. McNulty, MD, Kaiser Permanente Medical Center, 2200 O’Farrell Blvd, San Francisco, CA 94115 (edward.j.mcnulty@kp.org).

Author Contributions: Drs McNulty and Hung had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: McNulty.

Acquisition of data: Hung, Almers.

Analysis and interpretation of data: McNulty, Hung, Go, Yeh.

Drafting of the manuscript: McNulty.

Critical revision of the manuscript for important intellectual content: McNulty, Hung, Almers, Go, Yeh.

Statistical analysis: McNulty, Hung, Almers.

Obtained funding: McNulty.

Administrative, technical, and material support: McNulty.

Study supervision: McNulty, Yeh.

Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Go reported receiving research grant support from Genentech. Dr Yeh reported receiving consulting fees from Kaiser Permanente Northern California Division of Research. No other disclosures were reported.

Funding/Support: This study was supported by a grant from the Kaiser Permanente Northern California Community Benefits Program.

Role of the Sponsors: Kaiser Permanente had no role in the design and conduct of the study; the collection, management, and analysis, and interpretation of the data; the preparation, review or approval of the manuscript; or the decision to submit the manuscript for publication.

Additional Contributions: We thank Ralph Brindis, MD (University of California, San Francisco, School of Medicine), Matthew Solomon, MD, PhD (Kaiser Permanente Oakland Medical Center), and John Spertus, MD, for critical, uncompensated review of the manuscript.