The Impact of a High-risk Psychosocial Assessment on Outcomes After Durable Mechanical Circulatory Support

Abstract

Patient adherence is vital to the success of durable mechanical circulatory support (MCS), and the pre-MCS assessment of adherence by the multidisciplinary advanced heart failure team is a critical component of the evaluation. We assessed the impact of a high-risk psychosocial assessment before durable MCS implantations on post-MCS outcomes. Between January 2010 and April 2018, 319 patients underwent durable MCS at our center. We excluded those who died or were transplanted before discharge. The remaining 203 patients were grouped by pre-MCS psychosocial assessment: high-risk (26; 12.8%) versus acceptable risk (177; 87.2%). We compared clinical characteristics, nonadherence, and outcomes between groups. High-risk patients were younger (48 vs. 56; p = 0.006) and more often on extracorporeal membrane oxygenation at durable MCS placement (26.9% vs. 9.0%; p = 0.007). These patients had a higher incidence of post-MCS nonadherence including missed clinic appointments, incorrect medication administration, and use of alcohol and illicit drugs. After a mean follow-up of 15.3 months, 100% of high-risk patients had unplanned hospitalizations compared with 76.8% of acceptable-risk patients. Per year, high-risk patients had a median of 2.9 hospitalizations per year vs. 1.2 hospitalizations per year in acceptable-risk patients. While not significant, there were more driveline infections over the follow-up period in high-risk patients (27% vs. 14.7%), deaths (27% vs. 18%), and fewer heart transplants (53.8% vs. 63.8%).The pre-MCS psychosocial assessment is associated with post-MCS evidence of nonadherence and unplanned hospitalizations. Attention to pre-MCS assessment of psychosocial risk factors is essential to optimize durable MCS outcomes.

Introduction

Patients with advanced heart failure (HF) despite optimal medical and electrical therapies suffer significant morbidity and mortality,¹ and mechanical circulatory support (MCS) offers increased quality of life and survival.² The 2013 International Society of Heart and Lung Transplantation Guidelines for MCS recommend that all patients should be screened for psychosocial risk factors, medical compliance, social support, and coping skills before MCS implantation.³ Lack of sufficient support and limited coping skills are relative contraindications to MCS in patients with a history of nonadherent behavior and patients who demonstrate an inability to comply with medical recommendations on multiple occasions should not receive MCS.³

At our institution, the pre-MCS psychosocial assessment is performed by experienced licensed clinical social workers. Social workers will deem patients acceptable or high-risk for nonadherence after MCS implantation. At times, despite a high-risk assessment by the social worker, patients will undergo MCS implantation. This may happen when the assessment is considered inadequate due to the time constraints brought on by a patient’s clinical instability or when the selection committee feels a high-risk patient requires a chance to demonstrate future adherence. In some situations, the MCS device may allow a long-term “trial of adherence” without use of the scarce resource of a donor heart.

Studies have demonstrated that the presence of pre-MCS psychosocial risk factors impacts post-MCS readmissions and adverse events although the relationship between pre-MCS psychosocial risk factors and post-MCS adherence has not been investigated.^4–8 As patients’ adherence to the medication regimen, physician instruction, and clinic appointments is essential to maximize quality of life and survival and avoid morbidity from MCS complications, the purpose of this study was to determine the impact of a pre-MCS high-risk psychosocial assessment on post-MCS adherence and outcomes.

Methods

Patient Population

We reviewed patients who underwent durable MCS at Cedars-Sinai Medical Center (CSMC) in Los Angeles, California, between January 2010 and April 2018. Durable MCS was defined as HeartMate 2 left ventricular assist device (LVAD), HeartMate 3 LVAD, HeartWare LVAD, total artificial heart, or other biventricular support (bilateral HeartWare/bilateral thoratec paracorporeal ventricular assist device). Patients who died were transplanted before hospital discharge or who transferred care were excluded from the subsequent analysis because information on post-MCS adherence and outcomes would not be available. We divided the remaining patients based on the pre-MCS psychosocial assessment: high-risk or acceptable risk.

The psychosocial assessment was performed by a licensed clinical social worker with experience in managing patients with advanced HF who are being evaluated for MCS or heart transplantation. The psychosocial assessment was based on the report of the patient and their caregivers. If information was unclear or conflicting, input regarding the patients’ psychosocial characteristics would also be obtained from the patient’s primary treating cardiologist. While different social workers performed the assessments over the 9 years comprised by the study, the assessments were standardized to address six areas: caregiver plan, drug use, alcohol use, tobacco use, mental health, and adherence, as detailed in Table 1. No patients underwent MCS implantation without a pre-MCS psychosocial assessment. The presence of any one of the high-risk factors described in Table 1 would result in a classification of a high-risk psychosocial assessment.

Data Collection

Each patient’s preoperative data (demographic, clinical, echocardiographic, and laboratory data) were abstracted from the medical records. We defined post-MCS nonadherence as any of the following: more than three missed clinical appointments, more than three missed laboratory assessments, more than three episodes of medication nonadherence, more than three episodes of not following physician recommendations, or any evidence of substance use (alcohol, tobacco, or illicit drugs including opiates, cocaine, methamphetamines, and marijuana). Thus, if a MCS patient had 2 missed clinic appointments and 1 missed laboratory assessment, they would not meet criteria for post-MCS nonadherence; a patient had to exhibit more than three episodes of a given nonadherence marker (clinical appointments, laboratory assessments, medication nonadherence, not following physician recommendations) post MCS to be assigned the outcome of post-MCS nonadherence.

Adverse events were collected, including renal failure, gastrointestinal bleeding, stroke, pump thrombosis, and infection. All MCS complications followed the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) definition. We also collected unplanned hospitalizations post MCS and death.

The study protocol was reviewed and approved by the institutional review board at CSMC.

Data Analysis

All authors had full access to all the data in the study and take responsibility for its integrity and the data analysis. Baseline data are presented as percent or median (interquartile range). Comparisons of characteristics between patients deemed high-risk versus acceptable risk on the pre-MCS psychosocial assessment were done by the Fisher’s exact test for categorical variables and the Mann–Whitney test for continuous variables.

To determine the impact of pre-MCS factors on post-MCS outcomes, we performed a multivariable logistic regression analysis using age, pre-MCS extracorporeal membrane oxygenation (ECMO) support, INTERMACS Profile 1 at MCS implant, use of biventricular MCS, and high-risk pre-MCS psychosocial assessment on evidence of post-MCS nonadherence as well as a Cox proportional hazards model of the impact of pre-MCS psychosocial risk factors on time to post-MCS nonadherence. Cox proportional hazards models of the impact of pre-MCS psychosocial assessment on time to driveline infection, transplant, and death were also performed.

A multivariable analysis of the impact of the pre-MCS psychosocial assessment on readmissions could not be performed by logistic regression as 100% of high-risk patients were readmitted and thus high-risk status predicted readmissions perfectly. Instead, a Cox proportional hazards model was used to determine the impact of a high-risk psychosocial assessment on time to first hospitalization. As patients had multiple readmissions, a multivariable negative binomial model was also performed to assess the impact of the pre-MCS psychosocial assessment on the incidence rate (number of hospital readmissions over a person day).

We performed a Hosmer and Lemeshow’s goodness-of-fit test to assess the goodness-of-fit of the logistic regression model of post-MCS nonadherence, which yielded a p = 0.85, indicating a good fit. The Supremum test of the Cox model of time to hospitalization, time to driveline infection, time to transplant, and time to death confirmed that the proportional hazards assumption held for each covariate.

Results

Demographic and Clinical Characteristics

Of the 319 patients who underwent durable MCS at CSMC between January 2010 and April 2018, 203 were discharged with MCS from the implant hospitalization and followed at CSMC. Patients who died (n = 62) were transplanted before hospital discharge (n = 48) or who transferred care (n = 5) were excluded from the subsequent analysis because information on post-MCS adherence and outcomes would not be available.

Of the 203 patients for whom outpatient post-MCS follow-up was available, 26 (12.8%) were high-risk on psychosocial assessment. Those deemed high-risk were younger (48 vs. 56 years; p = 0.006) and were more often self-reported Hispanic ethnicity (34.6% vs. 18.1%; p = 0.03). There were no significant differences in other demographic characteristics, including sex, race, insurance, or level of education between those deemed high versus acceptable risk by psychosocial assessment (Table, Supplemental Digital Content 1, http://links.lww.com/ASAIO/A523).

Regarding clinical characteristics, groups did not differ in ejection fraction, left ventricular dimensions, cause of HF, presence of diabetes, or chronic kidney disease (Table, Supplemental Digital Content 1, http://links.lww.com/ASAIO/A523). The preimplant strategy was the same, with the majority of both groups having devices placed as bridge to transplant. There was a higher proportion of patients with high psychosocial risk who were on ECMO at the time of MCS placement, 26.9% vs. 9.0%; p = 0.007). Notably, patients with a higher pre-MCS psychosocial risk assessment also had shorter median time between psychosocial assessment and MCS implantation: 14 versus 38.5 days (p = 0.03), suggestive of more urgent nature of implantation due to clinical instability.

Psychosocial Risk Factors

Of the 26 patients deemed high-risk by psychosocial assessment, 11 (42.3%) had inadequate caregiver support, 15 (57.7%) had a history of nonadherence, 10 (38.5%) had active use of alcohol, tobacco, or illicit drugs. The proportion of patients with 1, 2, or 3 of these high-risk markers is shown in Figure 1.

Figure 1.

Reasons for pre-MCS high-risk psychosocial assessment. MCS, mechanical circulatory support.

Post-MCS Nonadherence

Patients deemed high-risk by pre-MCS psychosocial assessment had greater evidence of post-MCS nonadherence, including a significantly higher incidence of greater than three missed clinic appointments (23.1% vs. 4.0%), greater than three missed laboratory checks (42.3% vs. 6.2%), greater than three episodes of incorrect medication administration (46.2% vs. 10.7%), greater than three episodes of nonadherence to physician instructions (53.8% vs. 12.4%), use of alcohol (7.7% vs. 1.7%) and illicit drugs (23.1% vs. 0.6%; Figure 2).

Figure 2.

Markers of post-MCS nonadherence by pre-MCS psychosocial risk category. Red bars represent patients with a high-risk pre-MCS psychosocial assessment. Blue bars represent patients with an acceptable pre-MCS psychosocial assessment. p < 0.001 for all comparisons except post-MCS tobacco use where p = 0.065. MCS, mechanical circulatory support.

The impact of pre-MCS factors on post-MCS nonadherence was assessed. A high-risk pre-MCS psychosocial assessment was the strongest predictor of post-MCS nonadherence (OR, 7.5; 95% confidence interval [CI], 3.1–18.0), and this relationship persisted after adjusting for possible confounders, age, pre-MCS ECMO support, INTERMACS Profile 1 at MCS implant, and use of biventricular MCS (OR, 8.0; 95% CI, 3.0–20.8).

We also investigated the impact of number of pre-MCS psychosocial risk factors on post-MCS nonadherence. There was no significant difference in the risk of post-MCS nonadherence: compared with no MCS psychosocial risk factors, the OR for 1, 2, and 3 pre-MCS psychosocial risk factors on post-MCS nonadherence was 1.5 (95% CI, 0.9–2.5), 2.3 (95% CI, 0.8–6.2), and 4.0 (95% CI, 0.3–45.2), respectively.

We investigated the relationship between pre-MCS high-risk psychosocial assessment and time to post-MCS nonadherence. Patients deemed high risk by pre-MCS psychosocial assessment had a significantly lower mean days to MCS nonadherence (121 vs. 207 days; p = 0.047). However, there was no significant difference in time to post-MCS nonadherence by Cox proportional hazards estimate.

Post-MCS Outcomes

After a mean follow-up of 15.3 months, 100% of the high-risk patients had an unplanned hospitalization compared with 76.8% of the acceptable-risk patients (p = 0.006; Figure 3). High-risk patients had a median of 2.9 hospitalizations per year versus 1.2 hospitalizations per year in acceptable-risk patients (p = 0.03), although the median days out of the hospital did not differ between groups (297 days for high-risk patients vs. 366 days for acceptable-risk patients; p = 0.65).

Figure 3.

Post-MCS outcomes by pre-MCS psychosocial risk category. Red bars represent patients with a high-risk pre-MCS psychosocial assessment. Blue bars represent patients with an acceptable pre-MCS psychosocial assessment. MCS, mechanical circulatory support.

There was no significant difference in the incidence of MCS complications, including renal failure requiring dialysis, gastrointestinal bleeding, pump thrombosis, or stroke. While not significant, there were more driveline infections more than the follow-up period in high-risk patients (27% vs. 14.7%), deaths (27% vs. 18%), and fewer heart transplants (53.8% vs. 63.8%). This difference was not significant in a Cox proportional hazards model: hazard ratio (HR) 0.98 for driveline infection (95% CI, 0.44–2.2; p = 0.98), HR 0.84 for time to transplant (95% CI, 0.50–1.4; p = 0.51), and HR 1.6 for time to death (95% CI, 0.70–3.6; p = 0.27).

The indications for hospitalizations were compared between groups (Table 2). Of 133 hospitalizations in 26 patients in the high-risk pre-MCS psychosocial assessment group, the major indications for hospitalizations were infection (28; 21%), HF (24; 18%), arrhythmias (18; 14%), and anticoagulation adjustment (17; 13%). Of 489 hospitalizations in 177 patients in the acceptable-risk pre-MCS psychosocial assessment group, the major indications for hospitalization were infection (103; 21%), bleeding (93; 19%), and anticoagulation adjustment 53 (11%). Patients with high-risk pre-MCS psychosocial profiles were more likely to be admitted for HF (18% vs. 7%; p = 0.0001) and less likely to be admitted for bleeding (8% vs. 19%; p = 0.003). Other indications for hospitalization did not differ between deemed high-risk versus acceptable risk by pre-MCS psychosocial assessment.

With regard to the impact of the number of pre-MCS psychosocial risk factors on post-MCS hospitalizations, there was no significant difference in the risk of post-MCS hospitalization: compared with no MCS psychosocial risk factors, the HR for 1 2, and 3 pre-MCS psychosocial risk factors on time for first post-MCS readmission was 1.5 (95% CI, 0.2–2.0), 2.4 (95% CI, 1.5–4.2), and 2.1 (95% CI, 0.7–6.7), respectively.

A multivariable Cox proportional hazards model of the impact of the pre-MCS psychosocial assessment on time to first readmission demonstrated that a high-risk pre-MCS psychosocial assessment was associated with a 2.6-fold higher risk of first readmission (HR, 2.6; 95% CI, 1.7–4.1; Figure 4). This relationship persisted after adjustment for age, pre-MCS ECMO support, INTERMACS Profile 1 at MCS implant, and use of biventricular MCS (HR, 2.9; 95% CI, 1.8–4.6).

Figure 4.

Time to readmission by pre-MCS psychosocial risk category. MCS, mechanical circulatory support.

A multivariable negative binomial model of the impact of the pre-MCS psychosocial assessment on the incidence rate (number of hospital readmissions over a person day) demonstrated that a high-risk pre-MCS psychosocial assessment is associated with a rate of readmission 1.7 (95% CI, 1.1–2.5) times that of an acceptable-risk pre-MCS psychosocial assessment. This positive association persisted after adjustment for age at implantation, pre-MCS ECMO support, INTERMACS Profile 1 at MCS implant, and use of biventricular MCS.

Sensitivity Analysis

Of the 310 patients who underwent MCS implantation at our center between 2010 and 2018, 62 died before hospital discharge. We sought to determine if the proportion of patients with a pre-MCS high-risk psychosocial assessment was higher in the group of patients who did not survive to hospital discharge. Of those patients who died before hospital discharge, 4 (6.2%) were deemed high-risk by pre-MCS psychosocial assessment, which was not significantly higher than the 26 (12.8%) of 203 patients with MCS who survived to hospital discharge (p = 0.14).

Because the likelihood of readmission and complexity of follow-up may not have been the same for LVAD, total artificial heart, and other biventricular support, we performed a separate analysis for patients with durable LVAD support alone. The results were comparable: LVAD patients with a high-risk pre-MCS psychosocial assessment had a 5.9-fold increased risk (95% CI, 2.1–16.7; p < 0.0001) of post-MCS nonadherence and a 2.8-fold increase risk (95% CI, 1.7–4.7; p < 0.0001) of time to first hospitalization.

Discussion

Durable MCS devices provide increased quality of life and survival for patients with end-stage HF, but the burden on patients and caregivers is great.⁹ Thus, it is essential to perform a pre-MCS psychosocial assessment of medical adherence and social support before embarking on this aggressive intervention. This study demonstrates that the pre-MCS assessment by experienced licensed clinical social workers provides an accurate assessment of post-MCS behavior: patients deemed high-risk are more likely to miss clinic appointments or required laboratory evaluations, take medications incorrectly, and fail to comply with physician instructions. Such patients are also more likely to use alcohol or illicit drugs after MCS implantation. The consequence of such behaviors on post-MCS outcomes is greater than twofold increased rate of unplanned post-MCS readmissions.

There is no specific definition of adherence in the MCS population.¹⁰ In general, adherence is the extent to which medication intake behavior corresponds with the recommendations of the health care provider.¹¹ Thus, we devised a post-MCS nonadherence endpoint that was clinically relevant to the MCS populations. As frequent clinic visits, laboratory checks, medication adjustments, and physician recommendations are essential components of post-MCS care, and because use of alcohol, tobacco, and illicit drugs is strictly forbidden, we felt that violation of these directives would be important indicators of post-MCS nonadherence. While occasional missed clinic visits or laboratory checks could occur due to illness, repeated instances of nonadherence would be unlikely to occur solely because of poor health or symptoms and thus we chose to strictly define MCS nonadherence as more than three instances of missed clinic visits, more than three missed laboratory checks, more than three medication errors, or more than three episodes of not following physician instructions. A patient had to exhibit more than three episodes of one of these factors, not a combination of factors, to be deemed nonadherent for that factor. Using this strict, clinically intuitive definition, patients deemed high-risk pre-MCS were still more likely to exhibit nonadherence post MCS.

Patients with a high-risk pre-MCS psychosocial assessment were younger, which would be expected as this demographic is at greater risk of illicit substance use which would confer a high-risk psychosocial assessment. Three times as many high-risk patients were on ECMO at the time of MCS implantation compared with patients of acceptable psychosocial risk. This is also expected, as such patients are in extremis with few options, and such patients are more likely to receive “the benefit of the doubt” in placing a durable MCS device without clear evidence of prior adherence or illicit substance use.

In an analysis of Medicare beneficiaries with LVADs, between 2004 and 2010, approximately 80% of patients were rehospitalized within 1 year of implant, consistent with the 76.8% of patients in the acceptable-risk group who were rehospitalized in our analysis.¹² The Center for Medicare Services payment for all rehospitalizations in the first year was $53,560 in 2010; the fact that 100% of patients in the high-risk group had unplanned readmissions after MCS placement is higher than this national average with higher associated costs. Unplanned hospitalizations place a high burden on the patient, caregivers, medical professionals, and health care system, and leads to a worsened quality of life for patients.

These results complement a number of prior studies on the impact of pre-MCS psychosocial assessments on post-MCS outcomes. Psychosocial assessment has been associated with more emergency room visits, urgent visits, and readmissions,^5,8 fewer out-of-hospital (outpatient) days after discharge,⁷ and adverse cardiac events.^4–13 Our study extends these findings as it is the first to demonstrate that patients with pre-MCS high-risk psychosocial characteristics not only are at higher risk for unplanned hospitalizations but are also at higher risk for post-MCS nonadherent behaviors.

There was no significant association between a high-risk psychosocial pre-MCS assessment and outcomes besides unplanned hospitalizations, including outcomes that might be associated with poor adherence such as driveline infections, pump thrombosis, or death. It is possible that with a longer follow-up period or with a large sample of a single MCS device, such an association may be observed. However, we felt it important in this analysis to maximize generalizability by including all patients with durable MCS implanted more than the 9 year period as adherence to a strict medical regimen is crucial to the success of all devices, although the sensitivity analysis indicated comparable results when only patients with LVADs were considered.

It is also possible that a larger sample size would allow the development of a psychosocial risk score that would determine which of factors or combination of factors from the psychosocial assessment (caregiver plan, drug use, alcohol use, tobacco use, mental health, or adherence) was associated with the greatest risk of post-MCS nonadherence or poor outcomes. However, the current sample size was too small to determine the relative detriment of psychosocial risk factors. In addition, in our clinical practice, they are weighted equally in that the presence of any one rendered the patient high-risk on the social workers’ assessment.

A limitation of this study is that the impact of standardized assessment tools such as the Transplant Evaluation Rating Scale,⁷ the Psychosocial Assessment of Candidates for Transplant scale,⁶ and the Stanford Integrated Psychosocial Assessment for Transplantation (SIPAT)⁴ on post-MCS outcomes was not evaluated. At our center, such standardized tools have not routinely been employed for evaluation of patients before implantation of durable MCS. However, we believe these results are still useful and valid as they support the utility of the social worker’s clinical psychosocial assessment. We anticipate that the SIPAT will be used in the future, which will allow further study comparing the impact of social worker assessment with and without the SIPAT on post-MCS outcomes.

Another limitation of this study is that we did not assess the outcomes of patients who were declined for durable MCS on the basis of a high-risk psychosocial assessment. This denominator is important to understand the full picture of risk in this advanced HF population. But we can extrapolate, based on natural history studies of ambulatory advanced HF patients on inotropic support, that survival is 50–70% at 3 months,^14,15 while in our population, for those patients who survived to hospital discharge, the 1 year survival in the group with a high-risk pre-MCS psychosocial assessment was 70%. The question then becomes, if imminent death is almost certain without MCS, what is the implication of a high-risk psychosocial assessment pre-MCS?

In our program, psychosocial red flags in the absence of medical contraindications do present a dilemma. We may be more likely to place a durable MCS device than proceed with transplantation, as the durable MCS device provides a trial period to assess adherence without the use of a scarce donor heart. In these cases, the MCS device would function as a bridge to candidacy rather than bridge to transplant and may functionally be destination therapy if adherence criteria were not met post durable MCS. These results support that practice, insofar as high-risk patients have comparable survival to those patients with acceptable pre-MCS psychosocial risk and undoubtedly better survival that such patients without MCS.

Ultimately, the utility of the psychosocial assessment may be in preparedness planning: (1) identifying patients who will require more attention post MCS with reminders to keep appointments and follow medical instructions; and (2) identifying patients at high risk of recidivism to illicit substance use so that preventative mechanisms may be planned pre-MCS (such as alcoholics anonymous or outpatient drug treatment support programs).

In conclusion, the pre-MCS psychosocial assessment by experienced licensed clinical social workers is an integral component of the evaluation process. Those patients deemed high risk are more likely to exhibit harmful nonadherent behaviors post MCS and are more likely to suffer unplanned readmissions. Thus, when evaluating end-stage HF patients for MCS implantation, the potential for future nonadherence and burden to the patients, caregivers, and healthcare system must be recognized and factors into the complex decision-making process.

Biographies

Meet the Authors

Jaime D. Moriguchi, MD is Medical Director of the Mechanical Circulatory Support Program at the Cedars-Sinai Heart Institute. After earning his bachelor’s degree at Stanford University, Dr. Moriguchi went on to earn his medical degree at the David Geffen School of Medicine at the University of California, Los Angeles (UCLA). He completed his internship and residency, serving as chief resident, and completed a cardiology fellowship also at UCLA. Prior to joining Cedars-Sinai, Dr. Moriguchi was Co-Director of the Clinical Heart Failure Program and Medical Director of the Mechanical Circulatory Support Program at the UCLA Medical Center. Dr. Moriguchi has published more than 93 articles on a wide variety of subjects related to cardiac disease, cardiac transplantation, and heart failure. These articles have been published in the Journal of Heart and Lung Transplantation as well as other peer-reviewed publications. He has co-authored six book chapters on heart transplantation, has authored over 161 research abstracts and lectures regularly on mechanical circulatory support.

Michelle Kittleson is Professor of Medicine at Cedars-Sinai, Director of Education in Heart Failure and Transplantation, and Director of Heart Failure Research at the Smidt Heart Institute. She graduated from Harvard College and received her medical degree from Yale University. She completed residency training at Brigham and Women’s Hospital and cardiology fellowship at Johns Hopkins, where she also received a PhD in Clinical Investigation. Dr. Kittleson is Deputy Editor of the Journal of Heart and Lung Transplantation, on Guideline Writing Committees for the American College of Cardiology/American Heart Association and is the Co-Editor-in-Chief for the American College of Cardiology Heart Failure Self-Assessment Program. Her essays have appeared in New England Journal of Medicine, Annals of Internal Medicine, and JAMA Cardiology and her poems have been published in JAMA and Annals of Internal Medicine.