Understanding the demand for healthcare services required by a population is a challenge faced by US health systems and health plans alike. Health systems that wish to recruit physicians to practice in their communities create medical staff development plans that examine the balance between the demand for healthcare services and the supply of physicians in the community.1
Similarly, health plans must ensure that they maintain adequate capacity in their physician networks, which they do by considering the average patient’s demand for services and the number of patients enrolled in the health plan.2 If a health plan evaluates the quality of care delivered by the facilities or physicians within its network, the breadth of evaluations it must conduct is influenced by the same factors that affect necessary network size.
When planning health system or network capacity, a simplified assumption is often made that all populations demand similar levels of service, and that the level of demand has not shifted over time.1 Nonetheless, it has been established by the Dartmouth Atlas of Health Care and other research initiatives that substantial geographic variation exists in the degree of health service utilization.3-5 Similarly, several articles have characterized temporal variation in the utilization of assorted healthcare services.6,7
Although the utilization of many healthcare services can potentially be characterized by using claims data, outpatient cardiac catheterization is of particular interest because it is relatively common and has the potential for utilization variation across subpopulations because of the level of demographic and geographic variation in heart disease. For example, although 764 deaths from heart disease occurred per 100,000 Minnesotans aged ≥65 years between 2014 and 2016, there were nearly twice as many deaths (1376 per 100,000) among Oklahomans aged ≥65 years.8 Therefore, heart disease does not affect Americans uniformly across the nation.
To help health systems and health plans ensure adequate access and capacity to cardiac catheterization, this study characterizes the overall trend in the utilization of cardiac catheterization from 2012 to 2018, as well as how utilization has varied across subpopulations over time. Unlike previous research that focused on publicly insured populations,9,10 this analysis concentrates on a population of patients with commercial or Medicare Advantage health plans whose utilization of cardiac catheterization was reviewed by a nondenial prior authorization program.
This program reviews incoming orders using a rule-based system, and then outreaches to ordering providers for educative, peer-to-peer conversations with specialist physicians if the information provided in the order does not justify approval. Approval is provided even in circumstances in which the peer-to-peer conversation does not lead to consensus on the appropriateness of the order.11 Because all orders for cardiac catheterization were reviewed by the same nondenial prior authorization program for appropriateness, the influence of local practice styles might have been attenuated, leaving primarily differences in patient needs to drive the contrasts in the utilization of catheterization.
Methods
In this longitudinal, retrospective study, we examined trends in the utilization of cardiac catheterization from 2012 to 2018. Trend analyses were conducted for the overall population, as well as for a series of subpopulations. To understand how patient characteristics interact to influence the use of cardiac catheterization, we conducted an additional cross-sectional analysis to explore the factors that predicted whether a patient utilized cardiac catheterization in 2018. Using the patient-level 2018 data set, the association between patient characteristics and the number of catheterizations received was also examined.
This study was reviewed by the Advarra Institutional Review Board (Pro00040065), and received an exemption from Institutional Review Board oversight.
Outpatient cardiac catheterization claims data covering care delivered from 2012 to 2018 to patients between the ages of 18 and 89 years were extracted from the database of a national healthcare organization offering commercial and Medicare Advantage health plans. During all the years considered in the study, the health plans operated a nondenial prior authorization program that examined orders for outpatient cardiac catheterization.11
The Current Procedural Terminology codes used to define cardiac catheterization are provided in Appendix Table A1. Each claim was linked to demographic data for the patient to whom it pertained using a health plan enrollment database and other sources.
Contemporaneous health plan enrollment data were extracted and were used to determine the characteristics of all the patients who were enrolled in the health plans during the time of the study, regardless of whether they had cardiac catheterization. Each patient contributed 1 patient-month to the health plans’ member base for each month that the patient was enrolled, to enable analyses to consider that some patients might not have been enrolled in the health plan for an entire calendar year. As patients enroll in and disenroll from health plans over the course of a year, the size of the overall population, as well as the size of each subpopulation, was reported in terms of patient-years. One patient-year represents 12 months of health plan membership by 1 patient. Alternatively, 2 patients each enrolled for only half of a year would collectively account for 1 patient-year.
The following characteristics were noted for each patient: sex, age-band (18-29, 30-39, 40-49, 50-59, 60-64, 65-69, 70-79, 80-89), median annual income in home zip code (<$40,000, $40,000-$80,000, or >$80,000), state obesity rate (very low, <25%; low, 25%-29.9%; medium, 30%-34.9%; high, ≥35%), and urbanicity (rural or urban). Patients aged 60 to 64 years were categorized separately from patients aged 65 to 69 years, because Medicare eligibility starts at age 65 years for most Americans.12
The average income in the patients’ zip codes was estimated using the 2013 to 2017 American Community Survey’s 5-year estimates of median income, which reported income in 2017 inflation-adjusted dollars.13 The average rate of obesity in each state was determined using the 2018 obesity prevalence rates from the Behavioral Risk Factor Surveillance System survey released by the Centers for Disease Control and Prevention (CDC).14 The urbanicity of patients’ zip codes was derived from a zip code mapping table developed by the Centers for Medicare & Medicaid Services.15
We created a second data set, which consisted of all patients who were continuously enrolled in their health plan in 2018, regardless of whether they had cardiac catheterization. Patients were linked to demographic data using the same techniques used for the main longitudinal analysis. The number of patients included in this data set was slightly smaller than in the longitudinal analysis because of the additional requirement of continuous health plan enrollment.
The annual cardiac catheterization utilization per 12,000 health plan member-months was calculated for the overall population, as well as for each subsegment (eg, males vs females). We prorated the patients’ participation in the health plan if they were not enrolled for the full year, because the claims data only represent the care that the patients received during the months that they were enrolled in the plan. These rates were then normalized by 2012 levels, so that utilization in 2012 was reported as 100%, and utilization in all subsequent years was reported as a percentage of the 2012 levels.
For the follow-up analysis of the determinants of utilization, the number of catheterizations received by each patient in 2018 was noted. An additional binary variable was generated to indicate whether each patient had no catheterizations in 2018, or if they had 1 or more catheterizations in 2018.
We plotted the trend lines for each subgroup. The lines depict normalized cardiac catheterization utilization relative to 2012 levels. Spearman’s rank-order correlation was calculated for each trend line to determine whether it depicted a significantly monotonic trend (ie, a trend that is either nondecreasing or nonincreasing). A second set of trend lines were plotted to depict the ratio of each subpopulation’s utilization of cardiac catheterization relative to the overall population’s utilization in each year.
Using the patient-level data from 2018, we performed a multivariate logistic regression to evaluate the association between a patient having received a cardiac catheterization in 2018 and each of the categorical variables used in the trend analyses, namely, sex, age-band, median income in the patient’s home zip code, the obesity rate in the patient’s state, and urbanicity.
The results of the logistic regression were reported as adjusted odds ratios (ORs), with 95% confidence intervals (CIs). A multivariate Poisson regression was also conducted to examine the association between the categorical variables and the number of cardiac catheterizations received among the patients who had 1 or more cardiac catheterizations in 2018.
Results
The overall enrollment for the health plans, as well as the enrollment of each of the subpopulations, are presented in Table 1. The size of the population analyzed grew from approximately 2.5 million patient-years in 2012 to approximately 3.5 million patient-years in 2018. In all years, a slight majority of the patients were female, and the majority of patients were aged ≥65 years. In each year, at least 66% of patients resided in zip codes with median annual incomes of $40,000 to $80,000, or with incomes that were unknown. Similarly, in each year, at least 66% of patients came from medium obesity states, and at least 79% lived in an urban zip code area (Table 1).
The trendlines for cardiac catheterization utilization, which were normalized to 2012 utilization, are displayed for the overall population and for each subpopulation in Figure 1. These trendlines depict how cardiac catheterization utilization has changed over time within each subpopulation, so that within-population changes in utilization can be examined. Overall, cardiac catheterization utilization in 2018 was at 96% of its 2012 level.
Visual inspection and Spearman’s rank-order correlation (P = .40) did not suggest a monotonic trend; utilization exceeded the 2012 level in 2013 and 2016, but was below the 2012 level in the other years.
Similarly, the trends for each of the subpopulations had Spearman’s rank-order correlation values that did not provide evidence to suggest monotonic trends. The trendline with the most dramatic change, which was the trendline for low obesity states, was not monotonic (P = .09); although cardiac catheterization utilization decreased from 2012 to 2016, it increased slightly in 2017, before declining again in 2018.
The ratio of the utilization of each subpopulation to the overall population’s level of utilization is shown in Figure 2. In contrast to Figure 1, which shows the change in cardiac catheterization utilization within each subpopulation over time, Figure 2 demonstrates the trends in relative utilization between the subpopulations and the overall population. For example, although cardiac catheterization utilization by patients from zip codes with a median income of more than $80,000 grew overall by 16%, as shown in Figure 1, this subpopulation had relatively low utilization, and went from a utilization rate that was 54% of the population average in 2012, to 65% of the population average in 2018, as shown in Figure 2.
As seen in Figure 2, men used more cardiac catheterization services than the overall population in all years, ranging from 122% of the average utilization in 2012 to 126% in 2018. Younger subpopulations tended to use lower levels of cardiac catheterization than older subpopulations. The degree of utilization increased with each age-band between the bands covering ages 18 to 64 years.
The age-band consisting of patients aged 65 to 69 years used proportionately less cardiac catheterization in each year than the age-band consisting of patients aged 60 to 64 years. Although patients aged 60 to 64 years used 152% of the population average in 2012 and 138% in 2018, patients aged 65 to 69 years used 132% of the population average in 2012 and 106% in 2018. Patients aged 70 to 79 years had a relatively higher level of cardiac catheterization utilization than patients aged 65 to 69 years, and patients aged 80 to 89 years had a relatively lower level of utilization.
Patients’ cardiac catheterization utilization varied markedly in accordance with the median income of the community in which they lived. Patients who had various local conditions had markedly different levels of cardiac catheterization utilization. The patients residing in zip codes with a 2017 median income below $40,000 used more catheterization than the population average: 132% as much as in 2012 and 121% as much as in 2018. Similarly, patients residing in zip codes with a median income above $80,000 used less catheterization than the population average: 54% of the average in 2012 and 65% of the average in 2018. Patients residing in zip codes with intermediate or unknown incomes had utilization that was near the population average in all years (Figure 2).
Meanwhile, the highest levels of catheterization were observed in states with high obesity rates, 133% of the overall average in 2012 and 150% in 2018, and the lowest levels of cardiac catheterization utilization were observed in states with low obesity rates, at 38% of the population average in 2012 and 13% of the population average in 2018. States with very low obesity rates had a higher rate of cardiac catheterization than low obesity states, but a lower rate than medium and high obesity states. Finally, patients residing in rural areas used more catheterization in all years, ranging from 147% of the population average in 2012 to 138% of the population average in 2018 (Figure 2).
An adjusted analysis using patient-level 2018 data showed that each of the factors examined was significantly associated with whether a patient had cardiac catheterization. The data set consisted of 2,919,742 patients, of whom 2,854,329 (97.8%) did not have a cardiac catheterization in 2018. As shown in Table 2, male sex, rural urbanicity, residence in a zip code with a median income below $40,000, and residence in a high obesity state were all significantly associated with catheterization.
Compared with patients aged 18 to 29 years, patients aged 70 to 79 years had the greatest increased odds (OR, 220.79; 95% CI, 142.42-342.29) of cardiac catheterization, whereas those aged 80 to 89 years had the second greatest increased odds (OR, 207.75; 95% CI, 133.97-322.15). Female sex, residing in an urban area, residing in a zip code with a median income of more than $80,000, and residing in a state with a very low, low, or medium obesity rate were all associated with significantly reduced odds of undergoing cardiac catheterization.
We conducted an additional multivariate Poisson regression analysis using 2018 data to examine the association between each of the factors and the number of catheterizations patients received. Patients in rural areas had significantly more (P <.025) catheterizations. Patients residing in a low obesity state (P <.05) or a medium obesity state (P <.01) had significantly fewer catheterizations.
Discussion
The findings from this analysis suggest that cardiac catheterization utilization across the overall study population and across most subgroups remained stable from 2012 to 2018. Studies of temporal trends in the use of catheterization are mixed, and their findings depend on the type of catheterization examined and the population studied. One analysis of inpatient catheterization showed that between 2003 and 2014, the use of inpatient diagnostic right heart catheterization increased by 104%.16 A study of temporal trends in diagnostic coronary angiography in a veteran population between 2009 and 2015 showed that the use of coronary angiography remained relatively stable over time.9
In the US Military Health System’s population, diagnostic catheterization utilization was found to have declined 7.9% between 2011 and 2016.10 Given that our study population consisted predominantly of patients aged ≥65 years (Table 1), it is likely that the changes in the population’s catheterization utilization over time is more similar to that of a veteran population than an active-duty military population.
Finally, a study of patients aged 65 to 95 years with fee-for-service Medicare health plans reported that left heart catheterization with the use of left ventriculography declined between 2012 and 2016.17 That study used a set of procedure codes that were only partially overlapping with those used in our analysis, and featured a slightly different population, because people aged <65 years were not included.17
The age and sex mix of the population in our study also affected the aggregate utilization of cardiac catheterization observed; men used approximately 20% more catheterizations than the population average, and women used approximately 20% less catheterizations than the population average. This finding is not surprising, because the number of cardiovascular disease–related deaths is higher for men aged ≥35 years than for women aged ≥35 years, at 511.5 per 100,000 people and 348.5 per 100,000 people, respectively, according to CDC statistics from 2016 to 2018.8 The average level of utilization by the population as a whole was more reflective of the cardiac catheterization utilization by women than by men, because the population enrolled in the health plans had a slight female majority (Table 1).
Although a previous study of patients having cardiac catheterization in response to uncomplicated first acute myocardial infarction showed that age is inversely associated with the use of cardiac catheterization, the findings of our study were less clear-cut.18 From the 18 to 29 years age-band through the 60 to 64 years age-band, cardiac catheterization utilization increased progressively with age in our study. The age-bands of patients aged >64 years each had lower levels of utilization than that of the 60 to 64 years age-band. Although our study did not account for health plan type, one possible explanation for the discontinuity is that Medicare Advantage eligibility begins at age 65 years for most individuals.12 Therefore, it is possible that differences in health plan designs might have been a factor in the differing degrees of utilization.
Health inequities between rural and urban areas, as well as access to care issues, may likewise influence cardiac care utilization. The CDC has reported that the prevalence of obesity among adults is significantly higher in rural areas than in urban areas (34.2% vs 28.7%, respectively), which suggests a higher potential need for catheterization in rural areas.19 However, one study reported that in New Jersey, patients who resided further from healthcare services were less likely to have cardiac revascularization than patients whose residences were closer to those services.20 It is possible that the higher level of obesity in rural areas had a stronger impact on utilization than lack of access to care for rural patients.
Income is another social determinant of health that may play a role in influencing an individual’s need for catheterization. The use of catheterization was above the population average in zip codes with median incomes <$40,000. Low household income has been previously strongly associated with heart disease, as well as with the risk factors related to it.21 In developed nations, such as the United States, income is inversely correlated with obesity, which in our study was associated with higher catheterization utilization.22
Collectively, the findings of our study suggest that personal characteristics such as age and sex, and environmental characteristics such as urbanicity, local income, and the level of obesity in a patient’s home state, may contribute to the likelihood that a patient will need cardiac catheterization. Although the environmental characteristics studied in our study may be social determinants of health in the context of cardiac catheterization during the period observed, it may be possible to someday decouple factors such as urbanicity and income from health outcomes. Further research is needed to determine which interventions are most effective in reducing the influence of these environmental factors.
Limitations
The patterns reflected in the findings of this analysis are representative of the population studied, but may not be reflective of the overall US population. The patients were not evenly distributed across the United States, because they all had health plans from a healthcare organization operating predominantly in the South. As shown in Table 1, the demographics of the cohort changed over time on multiple dimensions, and had proportionally fewer people aged 18 to 29 years and proportionally more people aged 80 to 89 years in 2018 than in 2012. Although these demographic changes can make the overall trends observed in the population difficult to disentangle, the subgroups examined are likely to be more homogenous over time than the population as a whole.
In addition, the population predominantly consisted of older individuals with Medicare Advantage health plans. No people with traditional Medicare were included. Patients with Medicaid insurance were only included if they were dually eligible for Medicare Advantage.
It is possible that there were confounding factors driving some of the univariate trends depicted in Figures 1 and 2. For example, 96% of the patients who were residing in high obesity states in 2018 were not residing in zip codes that had a median income of more than $80,000. Nonetheless, the 2018 patient-level analysis (Table 2) considered these factors concurrently. Local income was significantly associated with the likelihood of catheterization, even after accounting for the obesity rate of a patient’s home state, and vice versa.
Conclusion
Health inequities between rural and urban areas, as well as access to care issues, may influence cardiac care utilization. Although the rate of cardiac catheterization was stable from 2012 to 2018 for the overall population and most subpopulations, there were marked differences in the rates of cardiac catheterization utilization between the subpopulations. Regardless of the year examined, there were consistent differences in utilization when patients were grouped by sex, age, local median income, state obesity rate, and urbanicity. To ensure adequate access to care, health plans and health systems should explore the implications of disproportionately high demand for cardiac catheterization in populations from lower-income areas, rural areas, and higher obesity rate states, as well as from older populations. The findings of this study suggest that demographic and environmental factors may influence an individual’s need for cardiac catheterization. Further research is needed to determine the reasons for these differences.
Author Disclosure Statement
The authors have no conflicts of interest to report.
References
- How many physicians make a health system? In: Finarelli HJ Jr. The Hospital Executive’s Guide to Physician Staffing. Marblehead, MA: HCPro; 2009:103-124.
- Kongstvedt PR. The provider network. In: Kongstvedt PR, ed. Essentials of Managed Health Care. 6th ed. Burlington, MA: Jones & Bartlett Learning; 2013:57-84.
- Song Y, Skinner J, Bynum J, et al. Regional variations in diagnostic practices. N Engl J Med. 2010;363:45-53. Erratum in: N Engl J Med. 2010;363:198.
- Welch HG, Sharp SM, Gottlieb DJ, et al. Geographic variation in diagnosis frequency and risk of death among Medicare beneficiaries. JAMA. 2011;305:1113-1118.
- The Dartmouth Institute for Health Policy and Clinical Practice. Tracking the care of patients with severe chronic illness. The Dartmouth Atlas of Health Care 2008. 2008. https://data.dartmouthatlas.org/downloads/atlases/2008_Chronic_Care_Atlas.pdf. Accessed July 30, 2021.
- Levin DC, Parker L, Halpern EJ, Rao VM. Recent trends in imaging for suspected coronary artery disease: what is the best approach? J Am Coll Radiol. 2016;13:381-386.
- Guichet PL, Duszak R Jr, Chaves Cerdas L, et al. Changing national Medicare utilization of catheter, computed tomography, and magnetic resonance extremity angiography: a specialty-focused 16-year analysis. Curr Probl Diagn Radiol. 2021;50:308-314.
- Centers for Disease Control and Prevention. Interactive Atlas of Heart Disease and Stroke. https://nccd.cdc.gov/DHDSPAtlas/Reports.aspx. Accessed March 31, 2020.
- Waldo SW, Gokhale M, O’Donnell CI, et al. Temporal trends in coronary angiography and percutaneous coronary intervention: insights from the VA Clinical Assessment, Reporting, and Tracking program. JACC Cardiovasc Interv. 2018;11:879-888.
- Eisenberg C, Mrsic Z, Cole G, et al. Recent trends in utilization of cardiovascular imaging and invasive coronary angiography in the large scale United States Military Healthcare System. J Am Coll Cardiol. 2018;71(11_suppl):A1637.
- Powell AC, Price SE, Nguyen K, et al. Prior authorization for elective diagnostic catheterization: the value of reviewers in cases with clinical ambiguity. Am Health Drug Benefits. 2018;11(4):168-176.
- Department of Health & Human Services. Who is eligible for Medicare? Updated September 11, 2014. www.hhs.gov/answers/medicare-and-medicaid/who-is-elibible-for-medicare/index.html. Accessed October 13, 2020.
- United States Census Bureau. S1903: Median income in the past 12 months (in 2017 inflation-adjusted dollars). 2017 American Community Survey 5-year estimates subject tables. https://data.census.gov/cedsci/table?q=S1903&tid=ACSST5Y2017.S1903. Accessed August 6, 2021.
- Centers for Disease Control and Prevention. Weight classification by Body Mass Index (BMI) (variable calculated from one or more BRFSS questions). 2018. https://nccd.cdc.gov/BRFSSPrevalence/rdPage.aspx?rdReport=DPH_BRFSS.ExploreByTopic&irbLocationType=StatesAndMMSA&islClass=&islTopic=&islYear=&rdRnd=62498. Accessed August 6, 2021.
- Centers for Medicare & Medicaid Services. Q4 2015 DMEPOS rural ZIP codes and revised CY 2016 fee schedule PUF formats. October 21, 2015. www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/DMEPOSFeeSched/DMEPOS-Fee-Schedule-Items/DME-Rural-Zip-and-Formats.html?DLPage=1&DLEntries=10&DLSort=2&DLSortDir=descending. Accessed October 4, 2019.
- Agarwal M, Patel B, Shah M, et al. National trends of diagnostic right heart catheterization in the United States. Circulation. 2017;136(suppl_1):Abstract 19490.
- Kini V, Viragh T, Magid D, et al. Trends in high- and low-value cardiovascular diagnostic testing in fee-for-service Medicare, 2000-2016. JAMA New Open. 2019;2:e1913070.
- Spencer FA, Goldberg RJ, Frederick PD, et al. Age and the utilization of cardiac catheterization following uncomplicated first acute myocardial infarction treated with thrombolytic therapy (the Second National Registry of Myocardial Infarction [NRMI-2]). Am J Cardiol. 2001;88:107-111.
- Centers for Disease Control and Prevention. CDC: more obesity in U.S. rural counties than in urban counties. June 14, 2018. www.cdc.gov/media/releases/2018/s0614-obesity-rates.html. Accessed October 12, 2020.
- Gregory PM, Malka ES, Kostis JB, et al. Impact of geographic proximity to cardiac revascularization services on service utilization. Med Care. 2000;38:45-57.
- Lemstra M, Rogers M, Moraros J. Income and heart disease: neglected risk factor. Can Fam Physician. 2015;61:698-704.
- Bentley RA, Ormerod P, Ruck DJ. Recent origin and evolution of obesity-income correlation across the United States. Palgrave Commun. 2018;4:146.