Abstract
Background:
Timely follow-up after cardiovascular hospitalization is recommended to monitor recovery, titrate medications, and coordinate care.
Objective:
To describe trends and disparities in follow-up after acute myocardial infarction (AMI) and heart failure (HF) hospitalizations.
Design:
Retrospective cohort study
Setting:
Medicare
Participants:
Medicare fee-for-service beneficiaries hospitalized between 2010 and 2019.
Measurements:
Receipt of a cardiology visit within 30 days of discharge. Multivariable logistic regression models were used to estimate changes over time overall and across five socio-demographic characteristics based on known disparities in cardiovascular outcomes.
Results:
The cohort included 1,678,088 AMI and 4,245,655 HF hospitalizations. Between 2010 and 2019, the rate of cardiology follow-up increased from 48.3% to 61.4% for AMI hospitalizations and from 35.2% to 48.3% for HF hospitalizations. For both conditions, follow-up rates increased for all subgroups, yet disparities worsened for Hispanic AMI patients and HF patients who were Asian, Black, Hispanic, Medicaid dual-eligible, and residents of counties with higher levels of social deprivation. By 2019, the largest disparities were between Black and White patients (AMI, 51.9% vs 59.8%, difference 7.9 pp [95% CI, 6.8–9.0]; HF, 39.8% vs 48.7%, difference 8.9 pp [95% CI, 8.2 – 9.7]) and Medicaid dual-eligible and non-dual eligible patients (AMI, 52.8% vs 60.4%, difference 7.6 pp [95% CI, 6.9 – 8.4]; HF, 39.7% vs 49.4%, difference 9.6 pp [95% CI, 9.2 – 10.1]). Differences between hospitals explained 7.3 pp [95% CI, 6.7 – 7.9] of the variation in follow-up for AMI and 7.7 pp [95% CI, 7.2 – 8.1]) for HF.
Limitations:
Generalizability to other payers.
Conclusions:
Equity informed policy and health system strategies are needed to further reduce gaps in follow-up care for AMI and HF patients.
Primary Funding Source:
National Institute on Aging
Following acute myocardial infarction (AMI) and heart failure (HF) hospitalization, rates of post-discharge adverse outcomes, such as mortality or hospital readmission, are high, and substantial socio-economic and demographic disparities exist (1–3). Over the past two decades, health policy initiatives, most notably the Centers for Medicare & Medicaid Services (CMS) Hospital Readmissions Reduction Program (HRRP), have focused on measuring and financially incentivizing improvements in post-discharge outcomes with mixed success (4,5).
Care coordination has been identified as a potentially important component of interventions to improve outcomes as patients transition from hospital to home (6,7). Patients face numerous coordination tasks including arranging follow up visits for laboratory tests, imaging, rehabilitation, and clinician office visits; initiating or changing cardiovascular medications; modifying home diet and exercise; and being alert for common complications. One potential avenue for improving post-discharge outcomes is early clinician follow-up to assist in care coordination, medication titration, and early detection of clinical deterioration. Early follow-up has been associated with high rates of evidence-based medication adherence and lower rates of readmissions.8−10 Recognizing this potential benefit, CMS implemented in 2013 a bundled payment program, known as transitional care management (TCM), which provides additional financial incentives to outpatient clinicians for managing the first 30 days of patients’ transition to the community after discharge from hospitals or nursing homes (7).
Contemporary data on rates of post-discharge follow-up are not well characterized. Prior studies of Medicare patients discharged from hospitals participating in the Get With the Guidelines-Heart Failure and the Acute Coronary Treatment and Intervention Outcomes Network Registry found that approximately half of patients received a post-discharge follow up within 14 days of discharge(9–11). However, these studies included hospitalizations between 2007 and 2011 that predated the HRRP, TCM, and other health-system level interventions, did not assess for inequities in follow-up rates, and may not be representative of all hospitals. Thus, we conducted a retrospective cohort study of all Medicare fee-for-service beneficiaries hospitalized for AMI or HF between 2010 and 2019 to characterize recent trends in timely post-discharge follow-up, whether disparities exist across demographic and socio-economic groups, and how disparities have changed during this time period.
METHODS
We analyzed Centers for Medicare and Medicaid Services (CMS) administrative inpatient and outpatient claims for 100% of traditional Medicare beneficiaries during the period 2010–2019 (including the first month of 2020 for follow-up). The cohort included all hospitalizations and observation stays for acute myocardial infarction (AMI) or heart failure (HF). AMI and HF admissions were defined using International Classification of Diseases codes used in CMS readmission and mortality measures (12,13).
To ensure full capture of outpatient care, we excluded beneficiaries who were not continuously enrolled in Medicare Parts A and B in the 12 months prior to hospital admission and 1 month following discharge (or until death if sooner), and beneficiaries enrolled in Medicare Advantage at any point during this same period. The unit of analysis was the facility stay; thus, beneficiaries could contribute multiple occurrences.
Given the study focus on ambulatory follow-up, hospitalizations resulting in death or discharge to hospice were excluded. Patients discharged to skilled nursing facilities were identified by overlapping claims (nursing home claim present on the date of hospital discharge or following day) and were included if beneficiaries were subsequently discharged to the community from the skilled nursing facility. For these episodes, the index date was the date of skilled nursing home discharge. Patients discharged to skilled nursing facilities who died, were directly readmitted without returning to the community, or were discharged to hospice were excluded.
Outcomes
The primary outcomes were 1) receipt of a post-discharge ambulatory cardiology visit within 30 days of discharge and 2) receipt of a post-discharge ambulatory cardiology or primary care visit within 30 days of discharge.
Ambulatory visits included evaluation and management, transitional care management, and preventative visits identified by Current Procedural Technology codes (99211–99215, 99201–99205, G0438, G0439, 99495–99496). Cardiologist visits included visits billed by general cardiologists and cardiology subspecialities. As Medicare does not report the specialty of advanced practice clinicians (APCs), we classified nurse practitioners and physician assistants who shared the same practice taxpayer identification number (TIN) as cardiologists as providing cardiology visits. Primary care visits included visits billed by physicians with specialties of general practice, family practice, internal medicine, osteopathic manipulative therapy, geriatrics, and preventive medicine, as well as visits billed by APCs who shared the same practice TIN as primary care physicians. Secondary outcomes included receipt of follow up by either a cardiology physician vs a cardiology APC within 30 days of discharge.
Socio-demographic variables
Given existing disparities in cardiovascular outcomes across demographic and socioeconomic groups, we sought to assess differential trends in post-discharge follow up rates for 5 socio-demographic groups as defined by sex, race, and ethnicity, Medicaid dual-eligibility, county-level social deprivation index (SDI), and rurality. Race and ethnicity were identified from the Medicare Beneficiary Summary File using the Research Triangle Institute race code and grouped as Asian/Pacific Islander, Black, Hispanic, White, or Other (including American Indian/Alaska Native and Unknown categories, due to low sample sizes) (14). The SDI is a composite measure of seven demographic characteristics collected at the county level in the American Community Survey which have been associated with poor health care access (15,16). The SDI is scored between 1 to 100 with higher scores reflecting more disadvantaged counties and was examined in quintiles in this analysis. Rurality was defined based on the 2010 rural-urban commuting area (RUCA) codes (17).
Statistical models included additional covariates of patient age, census region, individual Elixhauser comorbidities (18), claim-based frailty index (19), receipt of a primary care visit in the prior year (a marker for access to care), hospital length of stay, and discharge disposition.
Statistical analysis
We calculated descriptive characteristics for the AMI and HF cohorts. For each cohort, annual outcomes were assessed graphically and absolute differences were calculated comparing 2010 to 2019. We also calculated the variation in overall hospital-level follow-up rates.
For each cohort, we then estimate separate multivariable logistic regression analyses for each of the 5 socio-demographic groups. Each model included the primary predictor, year, an interaction between the primary predictor and year, and the other socio-demographic characteristics and covariates previously noted. Models included robust standard errors clustered by discharge hospital. To improve interpretability, we report average marginal effects instead of odds ratios to compare predicted probabilities of receiving cardiology follow-up over time and between socio-demographic groups. Identical models were constructed for the secondary outcome of receipt of cardiology or primary care follow-up within 30 days of discharge.
We sought to partition the observed socio-demographic differences into within- versus between-hospital components. First, we calculated quantile summary statistics to characterize the distribution of hospital-level follow-up rates. Second, we fit mixed effect logistic regressions to outcomes following hospitalizations in 2019, separately for each cohort. These models included hospital-level means of each socio-demographic characteristic (e.g., the hospital-level proportion of female patients) and the individual values of each socio-demographic variable, centered around the corresponding hospital mean (e.g., a covariate for the difference between the individual value of female indicator variable and the hospital-level proportion of female patients). The coefficients on these two terms allow us to decompose the effects into between- and within-hospital components (20,21). Third, to quantify the overall variation explained at the patient and hospital levels, we estimated variance partition coefficients (VPCs) from these models (22).
All analyses were conducted using Stata v17, with statistical significance assessed using 95% confidence intervals with α of .05. This project was approved by the Centers for Medicare & Medicaid Services privacy board and the Harvard Medical School institutional review committee, which also waived the requirement for obtaining informed consent because all administrative claims data were deidentified.
RESULTS
There were 1,678,088 Medicare fee-for-service hospitalizations discharges for AMI and 4,245,665 hospitalizations for HF included in the cohort. For the AMI cohort, 12.8% were younger than age 65 and 36.5% over age 80, 45.1% were female, 81.7% were white, 9.4% were Black, and 5.3% were Hispanic. For the HF cohort, 12.9% were under age 65 and 48.1% over age 80, 52.3% were female, 75.0% were White, 16.3% were Black, and 5.6% were Hispanic (Supplement Table 1).
Follow-up after cardiovascular hospitalizations
Between 2010 and 2019, the rate of cardiology follow-up within 30 days of discharge from an AMI hospitalization increased from 48.3% to 61.4%. A similar proportion of patients were seen by primary care as seen by cardiology and receipt of either cardiology or primary care follow-up increased from 70.6% to 79.7% (Figure 1A). Cardiology APC visits accounted for 3.0% of post-discharge visits in 2010 and 14.8% in 2019 (Figure 1B).
Figure 1.
Trends in 30-Day Follow-Up After Acute Myocardial Infarction Hospitalizations
A) Cardiology and Primary Care Follow-Up after Acute Myocardial Infarction Hospitalizations
B) Cardiology Follow-Up after Acute Myocardial Infarction Hospitalizations, by Clinician Type
Note: Cardiology clinic follow-up includes visits with cardiologist or advanced practice clinician (APC) linked to cardiology clinic based upon practice taxpayer identification number. Primary care follow-up includes visits with primary care physician or advanced practice clinician linked to primary care clinic. If patient was seen by both cardiologist and APC within 30-days classified as cardiologist visit.
Throughout the study period, follow-up after HF hospitalization was lower than follow-up after AMI hospitalization. Between 2010 and 2019, the rate of cardiology follow-up within 30 days of discharge from an HF hospitalization increased from 35.2% to 48.3% and receipt of either cardiology or primary care follow-up increased from 64.0% to 71.2% (Figure 2A). Throughout the study period, more HF patients were seen by primary care than by cardiology. Cardiology APC visits accounted for 2.4% of post-discharge visits in 2010 and 13.1% in 2019 (Figure 2B).
Figure 2.
Trends in 30-Day Follow-Up After Heart Failure Hospitalizations
A) Cardiology and Primary Care Follow-Up after Heart Failure Hospitalizations
B) Cardiology Follow-Up after Heart Failure Hospitalizations, by Clinician Type
Note: Cardiology clinic follow-up includes visits with cardiologist or advanced practice clinician (APC) linked to cardiology clinic based upon practice taxpayer identification number. Primary care follow-up includes visits with primary care physician or advanced practice clinician linked to primary care clinic. If patient was seen by both cardiologist and APC within 30-days classified as cardiologist visit.
Among patients receiving cardiology follow up, 7.1% of AMI and 11.0% of HF patients subsequently experienced readmission or mortality in the 30-day post-discharge period (Supplement Table 2).
Disparities in follow-up after acute myocardial infarction hospitalization
After adjustment for case-mix, predicted probabilities of cardiology follow-up after AMI hospitalization increased for all demographic and socio-economic groups between 2010 and 2019 (Table 1). The disparity in follow-up narrowed modestly for female compared to male patients (disparity in predicted probability difference for females 1.9 pp [95% CI, 1.4 – 2.4] in 2010 vs 0.6 pp [95% CI, 0.1 – 11] in 2019) and rural compared to urban patients (disparity in predicted probability difference for rural patients 9.6 pp [95% CI, 8.4 – 10.7] in 2010 vs 5.8 pp [95% CI, 4.7 – 7.0] in 2019). In contrast, disparities worsened for Hispanic patients (disparity in predicted probability 2.2 pp [95% CI, 0.8 – 3.6] in 2010 vs 5.6 pp [95% CI, 4.2 – 7.0] in 2019). Differences for Asian, Black, Medicaid dual-eligible patients as well as patients residing in counties with higher levels of social deprivation were similar in 2010 and 2019. By 2019, the largest disparities were between Black compared to White patients (51.9% vs 59.8%, difference 7.9 pp [95% CI, 9.0 – 6.8]) and Medicaid dual eligible and non-dual eligible patients (52.8% vs 60.4%, difference 7.6 pp [95% CI, 6.9 – 8.4]. Trends were largely similar for the secondary outcome of 30-day cardiology or primary care follow-up (Supplement Table 3).
Table 1.
Changes in 30-Day Cardiology Follow-Up After Acute Myocardial Infarction Hospitalization, by Demographic and Socioeconomic Groups
| Predicted Probability of 30-Day Cardiology Follow Up, % | Between-Group Differences, % | ||||
|---|---|---|---|---|---|
| 2010 | 2019 | Absolute Change | 2010 | 2019 | |
| Overall | 45.5 (44.8 – 46.3) | 58.5 (57.8 – 59.2) | 13.0 (12.3 – 13.7) | - | - |
| Sex | |||||
| Female | 44.3 (43.5 – 45.1) | 58.4 (57.6 – 59.1) | 14.1 (13.3 – 14.9) | −1.9 (−2.4 – −1.4) | −0.6 (−1.1 – −0.1) |
| Male | 46.2 (45.4 – 47.0) | 58.9 (58.2 – 59.7) | 12.7 (11.9 – 13.5) | Reference | Reference |
| Race/Ethnicity | |||||
| Asian/Pacific Islander | 43.5 (41.4 – 45.7) | 56.3 (54.3 – 58.2) | 12.7 (10.2 – 15.3) | −2.8 (−4.9 – −0.6) | −3.6 (−5.4 – −1.7) |
| Black | 38.5 (37.4 – 39.6) | 51.9 (50.7 – 53.1) | 13.4 (12.0 – 14.9) | −7.8 (−8.9 – −6.7) | −7.9 (−9.0 – −6.8) |
| Hispanic | 44.1 (42.7 – 45.5) | 54.2 (52.8 – 55.6) | 10.1 (8.3 – 11.8) | −2.2 (−3.6 – −0.8) | −5.6 (−7.0 – −4.2) |
| Other | 41.9 (39.5 – 44.2) | 57.5 (55.8 – 59.3) | 15.7 (13.0 – 18.4) | −4.4 (−6.7 – −2.1) | −2.3 (−3.9 – −0.7) |
| White | 46.3 (45.5 – 47.1) | 59.8 (59.1 – 60.6) | 13.5 (12.7 – 14.3) | Reference | Reference |
| Rurality | |||||
| Rural | 38.4 (37.2 – 39.7) | 54.4 (53.2 – 55.7) | 16.0 (14.8 – 17.2) | −9.6 (−10.7 – −8.4) | −5.8 (−7.0 – −4.7) |
| Urban | 48.0 (47.3 – 48.8) | 60.3 (59.6 – 61.0) | 12.3 (11.5 – 13.0) | Reference | Reference |
| Medicaid Dual Eligibility | |||||
| Dual Eligible | 39.7 (38.9 – 40.6) | 52.8 (51.9 – 53.7) | 13.1 (12.2 – 14.0) | −7.6 (−8.2 – −6.9) | −7.6 (−8.3 – −6.9) |
| Non-Dual Eligible | 47.3 (46.5 – 48.1) | 60.4 (59.7 – 61.1) | 13.1 (12.4 – 13.9) | Reference | Reference |
| County Level SDI Score | |||||
| 1–25 | 47.2 (46.1 – 48.4) | 61.1 (60.1 – 62.0) | 13.8 (12.7 – 14.9) | Reference | Reference |
| 25–50 | 46.0 (44.9 – 47.2) | 59.1 (58.0 – 60.2) | 13.1 (11.9 – 14.2) | −1.2 (−2.6 – 0.2) | −2.0 (−3.2 – −0.8) |
| 51–75 | 43.4 (42.2 – 44.6) | 57.5 (56.4 – 58.7) | 14.1 (12.9 – 15.3) | −3.8 (−5.3 – −2.3) | −3.5 (−4.9 – −2.2) |
| 76–100 | 44.4 (43.4 – 45.5) | 56.5 (55.5 – 57.5) | 12.1 (10.9 – 13.3) | −2.8 (−4.2 – −1.4) | −4.5 (−5.8 – −3.3) |
SDI, social deprivation index.
Note. Analysis restricted to hospitalizations of patients discharged alive and conducted using separate multivariable logistic regression models for each socio-demographic category with robust standard errors clustered by discharge hospital. Each model included the primary socio-demographic predictor, year, an interaction between the primary predictor and year, the other socio-demographic characteristics, and additional case-mix and hospitalization characteristics detailed in Supplement 1. Post-estimation average marginal effects were calculated to compare predicted probabilities of receiving cardiology follow-up over time and between socio-demographic groups.
Disparities in follow-up after heart failure hospitalization
Similarly, predicted probabilities of cardiology follow-up after HF hospitalization increased for all demographic and socio-economic groups between 2010 and 2019 and disparities for female and rural patients similarly narrowed. However, disparities were larger for Asian/Pacific Islander, Black, and Hispanic patients and worsened from 2010 to 2019 (Table 2). Disparities also worsened for Medicaid-dual eligible patients (disparity in predicted probability 8.0 pp [95% CI, 7.6 – 8.4] in 2010 vs 9.6 pp [95% CI, 9.2 – 10.1] in 2019) as well as patient residing in counties with higher levels of social deprivation (disparity in predicted probability comparing highest to lowest social deprivation quartile 3.5 pp [95% CI, 2.5 – 4.4] in 2010 vs 7.3 pp [95% CI, 6.3 – 8.3] in 2019). By 2019, the largest disparities were between Black and White patients (39.8% vs 48.7%, difference 8.9 pp [95% CI, 8.2 – 9.7]) and Medicaid dual eligible and non-dual eligible patients (39.7% vs 49.4%, difference 9.6 pp [95% CI, 9.2 – 10.1]). Trends were largely similar for the secondary outcome of 30-day cardiology or primary care follow-up though disparities for rural patients were not observed in 2019 (Supplement Table 4).
Table 2.
Changes in 30-Day Cardiology Follow-Up After Heart Failure Hospitalization, by Demographic and Socioeconomic Groups
| Predicted Probability of 30-Day Cardiology Follow Up, % | Between-Group Differences, % | ||||
|---|---|---|---|---|---|
| 2010 | 2019 | Absolute Change | 2010 | 2019 | |
| Overall | 32.1 (31.6 – 32.5) | 46.9 (46.4 – 47.3) | 14.8 (14.4 – 15.2) | - | - |
| Sex | |||||
| Female | 30.0 (29.5 – 30.4) | 45.6 (45.1 – 46.1) | 15.7 (15.2 – 16.2) | −4.0 (−4.3 – −3.7) | −2.0 (−2.3 – −1.7) |
| Male | 34.0 (33.5 – 34.5) | 47.6 (47.1 – 48.2) | 13.6 (13.1 – 14.2) | Reference | Reference |
| Race/Ethnicity | |||||
| Asian/Pacific Islander | 31.1 (29.5 – 32.6) | 41.1 (39.7 – 42.4) | 10.0 (8.2 – 11.9) | −2.3 (−3.9 – −0.8) | −7.6 (−8.9 – −6.3) |
| Black | 26.4 (25.8 – 27.1) | 39.8 (38.9 – 40.6) | 13.3 (12.5 – 14.1) | −6.9 (−7.6 – −6.3) | −8.9 (−9.7 – −8.2) |
| Hispanic | 28.9 (28.0 – 29.9) | 41.2 (40.2 – 42.2) | 12.3 (11.2 – 13.3) | −4.5 (−5.4 – −3.5) | −7.5 (−8.4 – −6.5) |
| Other | 29.0 (27.4 – 30.5) | 43.1 (41.7 – 44.4) | 14.1 (12.3 – 15.9) | −4.4 (−6.0 – −2.9) | −5.6 (−6.9 – −4.4) |
| White | 33.4 (32.9 – 33.9) | 48.7 (48.2 – 49.2) | 15.3 (14.8 – 15.8) | Reference | Reference |
| Rurality | |||||
| Rural | 26.4 (25.8 – 27.0) | 42.2 (41.4 – 43.0) | 15.9 (15.1 – 16.6) | −7.3 (−8.0 – −6.6) | −5.7 (−6.5 – −4.8) |
| Urban | 33.6 (33.1 – 34.1) | 47.9 (47.4 – 48.4) | 14.3 (13.8 – 14.8) | Reference | Reference |
| Medicaid Dual Eligibility | |||||
| Dual Eligible | 26.4 (25.9 – 26.9) | 39.7 (39.2 – 40.3) | 13.3 (12.7 – 13.9) | −8.0 (−8.4 – −7.6) | −9.6 (−10.1 – −9.2) |
| Non-Dual Eligible | 34.4 (33.9 – 34.9) | 49.4 (48.9 – 49.8) | 14.9 (14.4 – 15.4) | Reference | Reference |
| County Level SDI Score | |||||
| 1–25 | 33.5 (32.8 – 34.3) | 49.4 (48.7 – 50.1) | 15.9 (15.1 – 16.6) | Reference | Reference |
| 25–50 | 33.3 (32.5 – 34.1) | 48.7 (47.9 – 49.5) | 15.4 (14.6 – 16.2) | −0.2 (−1.2 – 0.8) | −0.7 (−1.7 – 0.3) |
| 51–75 | 30.6 (29.9 – 31.3) | 46.0 (45.1 – 46.8) | 15.3 (14.5 – 16.2) | −2.9 (−3.8 – −2.0) | −3.5 (−4.5 – −2.5) |
| 76–100 | 30.1 (29.4 – 30.8) | 42.1 (41.3 – 42.9) | 12.0 (11.2 – 12.8) | −3.5 (−4.4 – −2.5) | −7.3 (−8.3 – −6.3) |
SDI, social deprivation index.
Note. Analysis restricted to hospitalizations of patients discharged alive and conducted using separate multivariable logistic regression models for each socio-demographic category with robust standard errors clustered by discharge hospital. Each model included the primary socio-demographic predictor, year, an interaction between the primary predictor and year, the other socio-demographic characteristics, and additional case-mix and hospitalization characteristics detailed in Supplement 1. Post-estimation average marginal effects were calculated to compare predicted probabilities of receiving cardiology follow-up over time and between socio-demographic groups.
Hospital variation in follow-up
Hospitals varied widely in their 30-day cardiology follow up rates (Supplement Figure). After AMI hospitalization, the median was 44.8%, and the 25th and 75th percentiles were 27.8% and 57.9%. The corresponding quantiles of the distribution of follow-up after HF hospitalization were 32.7%, 22.9%, and 43.3%. We observed similar variation for primary care or cardiology follow up (Supplement Table 5).
For both cohorts, analyses decomposing socio-demographic characteristic effects into within-hospital and between-hospital differences, demonstrated that within-hospital difference remained for Black race, Hispanic ethnicity, rural residence, Medicaid dual-eligibility and county-level SDI score (Supplement Tables 6 and 7). Additionally, for HF but not AMI hospitalizations, within-hospital differences remained for female sex and all racial and ethnic groups. Between hospital-effects were largest for Medicaid dual-eligibility, for example, a 20% increase in the hospital-level the proportion of Medicaid dual-eligible patients was associated with a 21% and 25% lower odds of cardiology follow-up after AMI hospitalization and HF hospitalization, respectively. More modest between-hospital effects were observed for female sex, rural residence and Black race after both AMI and HF hospitalizations and for Other race and the counties with highest quartile of social deprivation. In the same models, examination of variation partition coefficients showed that 7.3% [95% CI, 6.7% – 7.9%] of the variation in follow-up after AMI hospitalization was attributable to hospitals and 7.7% [95% CI, 7.2% – 8.1%]) of the variation in follow-up after HF hospitalization, with the remainders attributable to individual variation.
DISCUSSION
This national longitudinal analysis of Medicare fee-for-service beneficiaries from 2010 to 2019 documents modest improvements in cardiology follow-up for patients discharged after an AMI or HF hospitalization. However nearly 40% of AMI and over 50% of HF patients were not seen by a cardiology clinician and nearly 20% of AMI and 30% of HF patients were not seen by any primary care or cardiology clinician within 30 days of discharge. Furthermore, while follow-up increased for all demographic and socio-economic groups, disparities in follow-up narrowed modestly for female and rural patients, but worsened for Asian, Black, Hispanic, and Medicaid dual-eligible patients, as well as patients residing in counties with higher levels of social deprivation.
This study updates prior estimates of small prospective cohort and registry studies of post-discharge follow-up from the 2000s, which similarly found lower rates of follow up for HF compared to AMI (9–11). In the subsequent decade and in a more representative sample, we find that 30-day cardiology follow up has improved to nearly 50% for HF patients and 60% for AMI patients. These findings have two important implications for future efforts to improve transitional care. First, transitional care for AMI and HF patients’ needs to be approached with a team-based model and cannot be thought of as solely the domain of cardiologists. This study finds that cardiology-based advanced practice clinicians are an increasing key part of post-discharge care and growth in timely cardiology follow up was primarily driven by increased APC visits. Additionally, many patients are seen in primary care and not cardiology in the post-discharge period. Many patients are likely to benefit from follow up with both cardiology and primary care given high rates of multimorbidity and polypharmacy, and thus strategies to coordinate post-discharge communication between patients, caregivers, primary care and cardiology is vital. Second, additional strategies need to be tested to increase rates of timely follow up. For instance, prior registry-based studies that have focused on whether or not a post-discharge cardiology appointment is scheduled have documented scheduled heart failure appointment rates of 50% to 80% - substantially higher than the rates we observed - suggesting there is a large implementation gap between appointment scheduling and actual completion of a visit (11, 23).
Timely follow-up may allow clinicians to assist AMI and HF patients in connecting with evidence-based care such as cardiac rehabilitation programs, monitor and treat early problems such as volume overload before they require emergency department care, and assess barriers to treatment adherence. The impact of timely routine post-discharge follow-up after AMI or HF hospitalization has not been formally tested in randomized trials. However earlier registry-based studies have demonstrated that timely follow up is associated with greater adherence to guideline-directed medical therapy after AMI and HF hospitalization and associated with lower readmission rates among HF patients (8–10). Further research using more recent and representative data is required to better understand the impact of timely follow up. Furthermore, prior randomized trials of enhanced transitional care programs, such as those involving more intensive nursing, home monitoring or home visits, have largely, but not exclusively, shown mortality benefits for HF patients (24–26) but have not been frequently studied in AMI patients (27). Given the rise of virtual visits during the COVID-19 pandemic, future research also should evaluate the use and outcomes of virtual visits for post-AMI and HF hospitalization follow-up, as this modality may increase access but it remains uncertain whether telehealth is adequate to assess post-hospital clinical status (e.g., volume status) (28).
The observed increase in timely post-discharge cardiology and primary care follow up was likely driven in part by Medicare policies enacted as part of the Affordable Care Act. Likely drivers of a focus on follow-up care include the HRRP which penalizes hospitals with higher-than-expected HF and AMI readmission rates and the TCM program which targets primary care clinicians and provides financial incentives timely follow-up and care coordination for all discharged patients (29). The combination of the positive financial incentive created by the TCM program and the financial penalty posed by HRRP to hospitals with higher-than-expected 30-day readmission rates effectively shifts CMS payments from low-performing hospitals to high-performing ambulatory clinics. This shift may, in part, explain the observed worsening demographic and socio-economic disparities in follow-up. Prior research indicates TCM adoption has been greater amongst larger practices and practices participating in accountable care organizations (30). Meanwhile the HRRP has led to a reduction in Medicare payments to safety-net hospitals and has been associated with widening of racial disparities in readmission rates within safety-net hospitals (2,31). Our findings of growing disparities in follow-up for Black, Hispanic and socio-economically disadvantaged groups, in conjunction with prior literature describing disparities in ambulatory care access broadly, further demonstrate the potential for unanticipated impacts of health policy on inequities (32,33). We also find a moderate proportion of variation of follow-up is explained by hospitals and that differences were driven by both within-hospital and between-hospital differences in outcomes across socio-demographics groups, with between-hospital effects larger for female, rural, and Medicaid dual-eligible patients. These findings add to the growing calls for explicitly incorporating equity into CMS policies and innovation models, such as the forthcoming, but voluntary, ACO Realizing Equity, Access, and Community Health (ACO REACH) model (34).
Additionally, during the same time period as our study, the initiation of voluntary value-based programs including the Medicare Shared Savings Program accountable care organizations and Bundled Payment for Care Initiative provided financial incentives to reduce spending and admissions for participating organizations that also may have led to an increasing focus on improving care transitions. Finally, meaningful use programs enacted in 2009 to encourage the adoption of electronic health records may have facilitated coordination and scheduling of follow-up visits (35). As these programs were rolled out in overlapping fashion throughout the early 2010s, untangling their effect is not possible, but prior research has found greater hospital participation in voluntary value-based programs was associated with greater reductions in 30-day readmissions (36).
This study has several limitations. First, patients may receive unbilled follow up, through phone calls or electronic messaging not captured by the current study. Second, we did not assess trends in post-discharge follow-up during the COVID-19 pandemic, which caused substantial disruptions in both inpatient and outpatient care delivery. Prior studies demonstrate there was substantial uptake of telehealth for the delivery of post-discharge ambulatory care during the early pandemic, though this did not impact overall follow-up rates (37,38). Whether post-hospital follow-up delivered by telehealth is associated with improved outcomes is an important avenue of future research. Third, though our study captures all Medicare fee-for-service beneficiaries, results may not be generalizable to commercially insured or Medicare Advantage populations. The observed patterns of care could, in part, be driven by changes in patient populations enrolled in traditional Medicare compared with Medicare Advantage, though recent studies suggest largely similar demographics and outcomes amongst AMI patients (38). Fourth, while a strength of our study was the inclusion of patients discharged to short-stay nursing homes and then home, we were not able to assess follow up for patients who permanently reside in long-term care nursing homes. Fifth, we did not assess the content of follow-up visits, such as whether additional diagnostic tests or medications were ordered. Sixth, analysis of racial groups are limited by validity of imputed race codes (14).
CONCLUSIONS
Between 2010 and 2019, timely post-discharge follow-up after AMI and HF hospitalizations increased substantially for Medicare fee-for-service beneficiaries. However nearly 40% of AMI and over 50% of HF patients were not seen by a cardiology clinician, and racial and socio-economic disparities have worsened. These findings indicate opportunities and challenges to equitably improving post-discharge care for all AMI and HF patients.
Supplementary Material
Acknowledgments
Sources of Funding:
Supported by a grant from the NIA (P01AG032952). Dr. Anderson was supported by career development awards from the American Heart Association (940950) and the National Institute on Aging (K76AG074878). Dr. Yeh was supported by a grant from the National Heart, Lung, and Blood Institute (K24HL150321). Dr. Marcantonio was supported by a grant from the National Institute on Aging (K24AG035075).
Disclosures:
Dr Anderson reported receiving grants from the American College of Cardiology, Boston OAIC Pepper Center, and US Deprescribing Research Network outside the submitted work. Dr Yeh reports research grants from Abbott Vascular, Abiomed, AstraZeneca, Cook, BD Bard, Boston Scientific, Medtronic, and Philips, all outside of the submitted work. Dr Yeh reports consulting for Abbott Vascular, AstraZeneca, Boston Scientific, and Medtronic. Dr. Landon reports receiving grants from the NIA and AHRQ and payments from RTI, Freedman Healthcare Consulting, and the ABIM, all outside of the submitted work.
Role of the Funder/Sponsor:
The American Heart Association, National Institute on Aging, and National Heart, Lung, and Blood Institute had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Footnotes
Disclaimer: The research reported in this publication was supported by the American Heart Association and the National Institute on Aging of the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the American Heart Association or the National Institutes of Health.
References
- 1.Krumholz HM, Normand ST, Wang Y. Twenty-Year Trends in Outcomes for Older Adults With Acute Myocardial Infarction in the United States. JAMA Netw Open. 2019. Mar 1;2(3):e191938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chaiyachati KH, Qi M, Werner RM. Changes to Racial Disparities in Readmission Rates After Medicare’s Hospital Readmissions Reduction Program Within Safety-Net and Non–Safety-Net Hospitals. JAMA Netw Open.2018;1(7):e184154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kyalwazi AN, Loccoh EC, Brewer LC, Ofili EO, Xu J, Song Y, Joynt Maddox KE, Yeh RW, Wadhera RK. Disparities in Cardiovascular Mortality Between Black and White Adults in the United States, 1999 to 2019. Circulation. 2022. Jul 19;146(3):211–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wadhera RK, Yeh RW, Joynt Maddox KE. The Hospital Readmissions Reduction Program - Time for a Reboot. N Engl J Med. 2019. Jun 13;380(24):2289–2291 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Cram P, Wachter RM, Landon BE. Readmission Reduction as a Hospital Quality Measure: Time to Move on to More Pressing Concerns? JAMA. 2022. Oct 25;328(16):1589–1590. [DOI] [PubMed] [Google Scholar]
- 6.Feltner C, Jones CD, Cené CW, Zheng ZJ, Sueta CA, Coker-Schwimmer EJ, Arvanitis M, Lohr KN, Middleton JC, Jonas DE. Transitional care interventions to prevent readmissions for persons with heart failure: a systematic review and meta-analysis. Ann Intern Med. 2014. Jun 3;160(11):774–84. [DOI] [PubMed] [Google Scholar]
- 7.Bindman AB, Blum JD, Kronick R. Medicare’s transitional care payment--a step toward the medical home. N Engl J Med. 2013. Feb 21;368(8):692–4 [DOI] [PubMed] [Google Scholar]
- 8.Hernandez AF, Greiner MA, Fonarow GC, et al. Relationship Between Early Physician Follow-up and 30-Day Readmission Among Medicare Beneficiaries Hospitalized for Heart Failure. JAMA. 2010;303(17):1716–1722. [DOI] [PubMed] [Google Scholar]
- 9.Daugherty SL, Ho PM, Spertus JA, et al. Association of Early Follow-up After Acute Myocardial Infarction With Higher Rates of Medication Use. Arch Intern Med. 2008;168(5):485–491. [DOI] [PubMed] [Google Scholar]
- 10.Faridi KF, Peterson ED, McCoy LA, Thomas L, Enriquez J, Wang TY. Timing of First Postdischarge Follow-up and Medication Adherence After Acute Myocardial Infarction. JAMA Cardiol. 2016;1(2):147–155. [DOI] [PubMed] [Google Scholar]
- 11.DeVore AD, Cox M, Eapen ZJ, Yancy CW, Bhatt DL, Heidenreich PA, Peterson ED, Fonarow GC, Hernandez AF. Temporal Trends and Variation in Early Scheduled Follow-Up After a Hospitalization for Heart Failure: Findings from Get With The Guidelines-Heart Failure. Circ Heart Fail. 2016. Jan;9(1):e002344. [DOI] [PubMed] [Google Scholar]
- 12.Krumholz HM, Lin Z, Drye EE, et al. An administrative claims measure suitable for profiling hospital performance based on 30-day all-cause readmission rates among patients with acute myocardial infarction. Circ Cardiovasc Qual Outcomes. 2011;4(2):243–252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Keenan PS, Normand SL, Lin Z, et al. An administrative claims measure suitable for profiling hospital performance on the basis of 30-day all-cause readmission rates among patients with heart failure. Circ Cardiovasc Qual Outcomes. 2008;1(1):29–37. [DOI] [PubMed] [Google Scholar]
- 14.Jarrín OF, Nyandege AN, Grafova IB, Dong X, Lin H. Validity of race and ethnicity codes in medicare administrative data compared with gold-standard self-reported race collected during routine home health care visits. Med Care 2020;58(1):e1–e8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Butler DC, Petterson S, Phillips RL, Bazemore AW. Measures of social deprivation that predict health care access and need within a rational area of primary care service delivery. Health Serv Res. 2013. Apr;48(2 Pt 1):539–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Social deprivation index (SDI). Robert Graham Center - Policy Studies in Family Medicine & Primary Care. (2018, November 5). Accessed March 10, 2023. https://www.graham-center.org/rgc/maps-data-tools/sdi/social-deprivation-index.html.
- 17.Loccoh EC, Joynt Maddox KE, Wang Y, Kazi DS, Yeh RW, Wadhera RK. Rural-Urban Disparities in Outcomes of Myocardial Infarction, Heart Failure, and Stroke in the United States. J Am Coll Cardiol. 2022. Jan 25;79(3):267–279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Moore BJ, White S, Washington R, Coenen N, Elixhauser A. Identifying Increased Risk of Readmission and In-hospital Mortality Using Hospital Administrative Data: The AHRQ Elixhauser Comorbidity Index. Med Care. 2017. Jul;55(7):698–705. [DOI] [PubMed] [Google Scholar]
- 19.Kim DH, Glynn RJ, Avorn J, Lipsitz LA, Rockwood K, Pawar A, Schneeweiss S. Validation of a Claims-Based Frailty Index Against Physical Performance and Adverse Health Outcomes in the Health and Retirement Study. J Gerontol A Biol Sci Med Sci. 2019; 74: 1271–1276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Fitzmaurice G, Laird N, Ware J. Applied Longitudinal Analysis, 2nd Edition. John Wiley & Sons; August 2011 [Google Scholar]
- 21.Begg MD, Parides MK. Separation of individual-level and cluster-level covariate effects in regression analysis of correlated data. Stat Med. 2003. Aug 30;22(16):2591–602. doi: 10.1002/sim.1524. [DOI] [PubMed] [Google Scholar]
- 22.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 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Figueroa JF, Wadhera RK, Frakt AB, et al. Quality of Care and Outcomes Among Medicare Advantage vs Fee-for-Service Medicare Patients Hospitalized With Heart Failure. JAMA Cardiol. 2020;5(12):1349–1357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Feltner C, Jones CD, Cené CW, et al. Transitional care interventions to prevent readmissions for persons with heart failure: a systematic review and meta-analysis. Ann Intern Med. 2014;160(11):774–784 [DOI] [PubMed] [Google Scholar]
- 25.Van Spall HGC, Rahman T, Mytton O, et al. Comparative effectiveness of transitional care services in patients discharged from the hospital with heart failure: a systematic review and network meta-analysis. Eur J Heart Fail. 2017;19(11):1427–1443. [DOI] [PubMed] [Google Scholar]
- 26.Van Spall HGC, Lee SF, Xie F, et al. Effect of Patient-Centered Transitional Care Services on Clinical Outcomes in Patients Hospitalized for Heart Failure: The PACT-HF Randomized Clinical Trial. JAMA.2019;321(8):753–761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Prvu Bettger J, Alexander KP, Dolor RJ, Olson DM, Kendrick AS, Wing L, Coeytaux RR, Graffagnino C, Duncan PW. Transitional care after hospitalization for acute stroke or myocardial infarction: a systematic review. Ann Intern Med. 2012. Sep 18;157(6):407–16. [DOI] [PubMed] [Google Scholar]
- 28.Wray CM. Post–Emergency Department Virtual Care—More Questions Than Answers. JAMA Netw Open.2022;5(10):e2237790. [DOI] [PubMed] [Google Scholar]
- 29.Agarwal SD, Barnett ML, Souza J, Landon BE. Medicare’s Care Management Codes Might Not Support Primary Care As Expected. Health Aff (Millwood). 2020. May;39(5):828–836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Anderson TS, Herzig SJ, Marcantonio ER, Yeh RW, Souza J, Landon BE. Medicare Transitional Care Management Program and Changes in Timely Postdischarge Follow-Up. JAMA Health Forum. 2024;5(4):e240417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Sheingold SH, Zuckerman R, Shartzer A. Understanding Medicare hospital readmission rates and differing penalties between safety-net and other hospitals. Health Aff (Millwood). 2016;35(1):124–131. [DOI] [PubMed] [Google Scholar]
- 32.Caraballo C, Ndumele CD, Roy B, et al. Trends in Racial and Ethnic Disparities in Barriers to Timely Medical Care Among Adults in the US, 1999 to 2018. JAMA Health Forum. 2022;3(10):e223856. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Mahajan S, Caraballo C, Lu Y, et al. Trends in Differences in Health Status and Health Care Access and Affordability by Race and Ethnicity in the United States, 1999–2018. JAMA. 2021;326(7):637–648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gondi S, Joynt Maddox K, Wadhera RK. “REACHing” for Equity - Moving from Regressive toward Progressive Value-Based Payment. N Engl J Med. 2022. Jul 14;387(2):97–99. [DOI] [PubMed] [Google Scholar]
- 35.Blumenthal D Wiring the health system: origins and provisions of a new federal program. N Engl J Med. 2011;365(24):2323–2329. [DOI] [PubMed] [Google Scholar]
- 36.Ryan AM, Krinsky S, Adler-Milstein J, Damberg CL, Maurer KA, Hollingsworth JM. Association Between Hospitals’ Engagement in Value-Based Reforms and Readmission Reduction in the Hospital Readmission Reduction Program. JAMA Intern Med. 2017;177(6):862–868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Anderson TS, O’Donoghue AL, Dechen T, Herzig SJ, Stevens JP. Trends in telehealth and in-person transitional care management visits during the COVID-19 pandemic. J Am Geriatr Soc. 2021. Oct;69(10):2745–2751. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Bressman E, Russo A, Werner RM. Trends in Outpatient Care and Use of Telemedicine After Hospital Discharge in a Large Commercially Insured Population. JAMA Health Forum. 2021;2(11):e213685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Landon BE, Anderson TS, Curto VE, Cram P, Fu C, Weinreb G, Zaslavsky AM, Ayanian JZ. Association of Medicare Advantage vs Traditional Medicare With 30-Day Mortality Among Patients With Acute Myocardial Infarction. JAMA. 2022. Dec 6;328(21):2126–2135. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.