Abstract
Objectives:
Increasingly, seriously ill Americans are ensured by Medicare Advantage (MA). However, standardized assessments of MA care quality do not highlight their unique experiences and needs. We adapted a previously validated survey to describe care experiences of seriously ill MA enrollees and highlight areas of unmet need for quality improvement.
Study Design:
This is a cross-sectional study.
Methods:
We sampled 3500 MA enrollees with the highest Hierarchical Condition Category (HCC) scores within one MA organization and administered a validated patient Serious Illness Survey for Home-Based Programs assessing communication, care coordination, symptom management, and planning for care in July and August 2023. We calculated top-box scores for all respondents and used linear regressions to calculate adjusted scores for subgroups of respondents meeting alternate serious illness definitions.
Results:
Of 1412 respondents, 84.6% had three or more chronic conditions. Responses indicated substantial room for improvement across quality measures assessed, with just 59.6% and 56.9% of respondents reporting that their communication and care coordination needs were met, respectively; 52.8% reporting that they “definitely” got the help they wanted for their symptoms; and 38.8% indicating that they had “definitely” discussed planning for care with their health care professionals. Respondents with HCC risk scores of ≥3 plus recent health care utilization also reported considerable unmet needs.
Conclusions:
Seriously ill MA enrollees have important unmet needs for care planning, care coordination, communication, and symptom management that are not highlighted by existing quality measures, but can be addressed through evidence-based interventions. Assessing care quality for seriously ill enrollees using standardized, survey-based measures can help to identify actionable improvement targets for this high-need, high-cost group.
Key Message
Seriously ill enrollees of a statewide MA organization reported substantial unmet needs for communication, care coordination, symptom management, and planning for care. Assessing care quality for seriously ill enrollees using validated measures can identify actionable improvement targets for this vulnerable, high-cost group.
Introduction
People with serious illness test a health care system’s ability to provide coordinated and responsive care that is aligned with their preferences and goals. 1 Increasingly, seriously ill Americans are insured by Medicare Advantage (MA): As of 2023, more than half of all Medicare beneficiaries were enrolled in MA, 2 and an estimated one-fifth had serious medical conditions. 3 Seriously ill MA enrollees have high needs for care, support services, and care coordination, and some could benefit from receiving some or all care in their homes. 4
Communication, care coordination, symptom palliation, and planning for care have been identified by expert stakeholders as key aspects of high-quality serious illness care. 1 These aspects of care can be measured directly by surveying patients and their family caregivers, and survey results can be used to identify opportunities for quality improvement and implementation of evidence-based interventions that have been proven to address specific quality gaps, such as communication programs, 5 care coordination initiatives, 6 and comprehensive, interdisciplinary palliative care.7–9
To date, limited research has assessed the care experiences of seriously ill MA enrollees. 10 Some prior studies rely on retrospective reports following enrollee death. For example, one recent study analyzed survey responses from bereaved family caregivers of MA enrollees and found that just two in five rated their family member’s care in the last month of life as “excellent.” While retrospective approaches are highly valuable for assessing the full episode of care at the end of life, prospective assessments can uniquely inform improvements to ongoing care.
Analyses of national patient surveys have found that MA enrollees in the last year of life and those with end-stage renal disease (ESRD) report similar or better experiences than all other enrollees11,12 and that MA enrollees with dementia report similar overall ratings of care but greater difficulty getting needed care than those without dementia.13,14 However, national initiatives to monitor and promote improvements in quality of care in MA have sample designs and content intended to assess care for all enrollees,15,16 rather than focus on those with serious illness and their specific needs, such as symptom management and care planning. 17
To address these gaps and allow for standardized and prospective assessment of serious illness care quality in MA, our team adapted a previously validated patient survey of serious illness care experiences. Quality measures from the original survey have been endorsed by the Centers for Medicare & Medicaid Services (CMS) consensus-based entity for use in quality reporting and value-based payment programs. 18 We administered the adapted survey to a sample of seriously ill MA enrollees in one statewide MA organization to highlight areas for quality improvement overall and across various definitions of serious illness.
Methods
Survey sample
We selected 3500 MA enrollees of Blue Cross Blue Shield of Massachusetts who were of age ≥18 years and who had at least one in-person health care visit (at home, an outpatient clinic, or an emergency department (ED), and inpatient and skilled nursing facility and rehab admissions) in the three months prior to sampling (February through April 2023). We used CMS’s Hierarchical Condition Category (HCC) risk scores to identify enrollees with serious illness. HCC risk scores combine information about MA enrollees’ demographics and diagnoses to summarize their complexity; risk scores are used to adjust MA payments and are often used to identify high-cost, high-need populations. We selected enrollees with the highest HCC risk scores of all Blue Cross Blue Shield of Massachusetts’ MA enrollees during that timeframe (average HCC risk score = 2.85, versus 0.59 for otherwise eligible enrollees) and enrollees with ESRD who are seriously ill but do not have risk scores assigned to them by the MA organization; enrollees were not selected if they were enrolled in hospice or requested not to be contacted for surveys. Selected enrollees were excluded if they died prior to the initiation of survey administration (n = 80) or did not read or speak English or Spanish (the languages in which the survey was being administered; n = 21).
Survey instrument
We adapted the previously validated Serious Illness Survey for Home-Based Programs 19 for use among MA enrollees with variable exposure to home-based care. Respondents were asked to assess care in the last three months from all health care professionals, including doctors, nurses, aides, social workers, case managers, and other people who provided care, including care received at their home, or at a doctor’s office, clinic, or hospital. The 44-item survey included evaluative items regarding communication, care coordination, symptom management, planning for care, and other topics of interest for internal quality improvement, as well as items about health and functional status, demographic characteristics, recent in-home and telehealth visits, and whether and how a proxy assisted with completion of the survey. Final full and abridged versions of the survey instrument, called the Serious Illness Survey for Community-Based Care, are available free online in English and Spanish. 20
Survey administration
We administered the survey in July and August 2023 via a mail–telephone protocol, which consisted of a prenotification letter, followed by a mail survey one week later, and up to five calls to complete the survey by phone if the mail survey was not returned after three weeks.
The mail survey cover letter and phone introductory script were addressed to the patient but indicated that a family member or friend could assist with or complete the survey for the patient if needed. The mail survey was available in English, with a Spanish-language note on the cover letter noting the availability of Spanish surveys. Spanish was offered as an option by telephone interviewers. The study was approved by the Institutional Review Board at RAND.
Analysis
We calculated response rates overall and by patient characteristics (sociodemographic characteristics, health status, and health care utilization patterns) and used logistic regression to examine the odds of responding to the survey from patient characteristics. We described characteristics of all respondents. In addition, because MA organizations often use claims- and electronic health record-based algorithms to identify seriously ill enrollees based on some combination of diagnoses, risk scores, and recent health care utilization, 3 we also describe characteristics of subgroups of respondents meeting three alternative definitions of serious illness. These alternatives build off of those proposed by national serious illness care initiatives 21 and are defined as follows: (1) all enrollees with HCC risk score of ≥3; (2) enrollees with HCC risk score of ≥2 plus an ED visit, inpatient hospitalization, or skilled nursing facility (SNF) stay in the last three months; and (3) enrollees with HCC risk score of ≥3 plus an ED visit, inpatient hospitalization, or SNF stay in the last three months.
For each evaluative item, we calculated the percentage of eligible respondents endorsing the most positive response option(s). This “top-box” scoring approach is used in public reporting of several national patient care experience surveys due to its ease of comprehension by consumers,22–24 and because it sets high standards for care quality. Respondents who endorse less positive response option[s] are described as having unmet needs. Only respondents eligible for a given evaluative item were included in score calculations (i.e., only respondents who indicate that they experienced and wanted help for pain are included in scoring of receipt of help for pain). Scored survey items were used to calculate four composite measures (communication, care coordination, help for symptoms, and planning for care), and one single-item measure (overall rating); development and psychometric properties of these measures are described elsewhere.19,25 Composite scores were calculated for those who responded to at least one item in the given composite, using centering to address missing data resulting from item nonresponse (i.e., skipped survey items).
We calculated unadjusted average item and composite scores for all respondents. We then used linear regressions to calculate scores for subgroups of respondents meeting the three serious illness definitions of interest, adjusting for patient age and education, as these characteristics are known to be associated with systematic differences in survey response. 26 We did not adjust for other variables recommended for case-mix adjustment of this survey (primary diagnosis, proxy response, self-reported functional status, or physical or mental health), as they may be conceptually similar to, or correlated or confounded with, the serious illness definitions. We imputed missing values for age and education with the overall mean across respondents; 27 missingness rates were low (0% and 2.8%, respectively).
Results
In all, 1412 MA enrollees or their caregivers responded to the survey, for an overall response rate of 41.5% (74.2% of responses by mail and 25.8% by phone). One-quarter (25.3%) of respondents had an HCC risk score of ≥3. Eighty-five percent of all respondents had three or more chronic conditions; 59.3% had diabetes, 46.7% had congestive heart failure, 42.9% had renal disease, 38.2% had cancer, and 9.3% had dementia (Table 1). Respondents meeting more stringent definitions of serious illness displayed different patterns of characteristics; for example, they were more likely than all respondents to report that they were in fair or poor physical health and that they received health care visits at home in the last three months (Table A1).
Serious Illness Survey Respondent Characteristics
Missing values not shown; between 0% and 3.6% of all respondents had missing values across characteristics.
Self-reported.
Count of HCC codes for the following chronic conditions: diabetes; neoplasms; liver disease; musculoskeletal disorder; blood disorder; psychiatric disorder; neurological disease; heart disease; cerebrovascular disease; vascular disease; eye disease; injury; dementia; CHF; COPD; metastatic cancer; lung and other severe cancers; breast, prostate and other cancers and tumors; and kidney disease or ESRD.
CHF, congestive heart failure; COPD, chronic obstructive pulmonary disorder; ED, emergency department; HCC, Hierarchical Condition Category; ESRD, end-stage renal disease.
Response rates did not vary significantly by HCC risk score; however, those with at least one ED visit, inpatient admission, or SNF stay in the last three months were significantly less likely to respond than those with none of these types of utilization (35.9% vs. 43.0%), those with dementia were significantly less likely to respond than those with other diagnoses (33.2% vs. 42.7%), those with more outpatient clinic visits in the last three months were significantly more likely to respond (43.0% among those with ≥6 visits versus 35.9% among those with zero or one visit), and those that received home visits of any kind in the last three months were significantly more likely to respond than those who did not (45.1% vs. 39.8%; Table A2).
Reported care experiences
Across all respondents, there was substantial room for improvement across quality measures assessed, with just 59.6% and 56.9% endorsing the most positive response options for the communication and care coordination composite measures, respectively (Table 2 and Table A3). Slightly more than half (52.8%) of respondents reported that they “definitely” got the help they wanted for their symptoms, and 38.8% of respondents indicated that they had “definitely” discussed planning for care with their health care professionals. Sixty-two percent of respondents rated their overall care a 9 or 10 out of 10.
Serious Illness Survey Measure and Item Top-Box Scores
A screening item and/or tailored nonapplicable response option (i.e., “I did not want…”) determines whether a respondent is eligible for this evaluative item.
With regard to specific aspects of care in greatest need of improvement, fewer than 40% of all respondents indicated that they “definitely” talked with health care professionals about how to get help with everyday activities, what their health care options would be if they got sicker, what was important to them in their lives, and the care or treatment they get from other doctors or health care providers. Fewer than 50% of respondents indicated that they “definitely” or “always” got the help they needed for their pain and anxiety or sadness.
Respondents reported more favorably on their care between visits, with 75.8% reporting that they “definitely” knew who to call if they had questions or needed help between visits, and 73.5% indicating that they “definitely” got the help they needed when they contacted their health care professionals between visits.
Respondents meeting more stringent definitions of serious illness (i.e., higher HCC risk scores plus acute health care utilization in the last three months) generally reported similar or slightly better care experiences, indicating substantial room for improvement regardless of how seriously ill enrollees were identified (Table 3). For example, those with an HCC risk score of ≥3 plus acute health care utilization in the last three months significantly more often indicated that they had “definitely” discussed planning for care with their health care professionals than all other respondents (47.1% vs. 38.3%); notably, however, even among the subgroup reporting better care planning experiences, fewer than half of respondents reported “definitely” having needed care planning discussions.
Serious Illness Survey Measure Top-Box Scores by Serious Illness Definition
Scores adjust for age and education.
Discussion
As more Medicare beneficiaries enroll in MA, assessing the quality of MA care is critical. This is particularly true for seriously ill enrollees, whose high needs place them at risk of under-delivery of needed care when organizations seek to reduce cost. 17 Although our respondents were enrolled in an MA organization with a 4-star rating from CMS (out of five possible stars), we found important opportunities to improve in each of the key domains of high-quality serious illness care.
Our results are highly actionable. We find specific and notable gaps in care planning, such as failure to ask about priorities for future care, that can be addressed by training primary care providers in best practices of advance care planning. 28 Visits from home-based serious illness programs designed to “wrap around” existing primary and specialty care may also increase conversations about care from other doctors and promote identification of those who need help with everyday activities like meals and transportation; in a prior study using the same survey, seriously ill individuals receiving care from a specialized home-based program reported substantially higher rates of discussions on these topics. 19 Referrals to interdisciplinary palliative care may also help to address the unmet needs of patients who report that they are not receiving sufficient help for their pain, breathing, or anxiety or sadness. 1
The seriously ill MA enrollees responding to our survey reported substantially poorer care experiences than MA enrollees responding to similar questions on an annual Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey. 29 This suggests that the care experiences of seriously ill enrollees differ notably from other MA enrollees.
Surveys of MA plan performance are designed to assess topics applicable to a high proportion of enrollees. 15 Topical domains that are primarily applicable to key subgroups, such as the seriously ill, are often omitted to limit survey length. 30 Moreover, an MA plan’s overall performance is not necessarily indicative of its performance for key subgroups.31,32 In a general survey, subgroup members also may be too few in number within any given plan to allow for reliable assessments of their care at the plan level. Drawing a targeted sample of seriously ill MA enrollees across plans and surveying them on aspects of care that are most important to them would enable stratified quality reporting. Stratified reporting that describes care of seriously ill MA enrollees can help to bring attention to their unique care needs.
We used a broad, easily replicable definition of serious illness to identify our overall survey sample. MA organizations often use narrower definitions to identify seriously ill enrollees for outreach and support given the costliness of these efforts. Notably—and perhaps counterintuitively—in our study, survey respondents meeting more stringent definitions of serious illness generally reported similar or slightly better care experiences than those meeting our general definition. This is perhaps because enrollees meeting the more stringent definitions were more likely to receive in-home visits that focus on care coordination and care planning.
Limitations
Our study uses a validated tool to assess a broad set of aspects of high-quality serious illness care. Nonetheless, it has some limitations. We collected survey data from living enrollees of one high-performing MA organization, which serves a mostly White, English-speaking population; therefore, results may not generalize to all seriously ill MA enrollees, particularly those in racial and ethnic groups underrepresented in our data, who typically report poorer care experiences;33,34 consequently, it is possible that unmet needs are even greater among some seriously ill MA enrollees than among our study population. Although we allowed for proxy respondents to answer on behalf of enrollees, response rates were still lower for enrollees with dementia diagnoses and certain types of recent health care utilization; therefore, our results may underrepresent the care experiences of those with dementia, who have unique care needs, and those with burdensome transitions across health care settings. As we use an observational design, we are not able to directly compare care quality between those who did and did not receive specialized serious illness care services, such as home-based palliative care.
Conclusions
Our survey of seriously ill MA enrollees identifies extensive unmet needs for communication, care coordination, symptom management, and planning for care that are not highlighted by existing quality measures, but can be addressed through evidence-based interventions. Assessing care quality for seriously ill enrollees using standardized, survey-based measures that assess key aspects of serious illness care can identify actionable improvement targets for this high-need, high-cost group.
Footnotes
Acknowledgments
The authors gratefully acknowledge the Medicare Advantage enrollees and family caregivers who participated in the study.
Author Disclosure Statement
No interests to disclose.
Funding Information
This work was supported by the Gordon and Betty Moore Foundation under grant agreement #6902.
Appendix
Serious Illness Survey Item Responses for Component Survey Items and Single-Item Measures
| Survey item | Response options | Response distribution among all respondents
a
|
|---|---|---|
| Communication | ||
| In the last three months, how often did your health care professionals spend enough time with you? | Never | 70 (5.1%) |
| Sometimes | 129 (9.4%) | |
| Usually | 400 (29.1%) | |
| Always | 777 (56.5%) | |
| In the last three months, how often did your health care professionals explain things to you in a way you could understand? | Never | 46 (3.3%) |
| Sometimes | 83 (6.0%) | |
| Usually | 361 (26.3%) | |
| Always | 885 (64.4%) | |
| In the last three months, how often did your health care professionals listen carefully to you? | Never | 41 (3.0%) |
| Sometimes | 95 (6.9%) | |
| Usually | 381 (27.9%) | |
| Always | 851 (62.2%) | |
| In the last three months, how often did you feel that your health care professionals cared about you as a whole person? | Never | 32 (2.3%) |
| Sometimes | 109 (7.9%) | |
| Usually | 403 (29.4%) | |
| Always | 829 (60.4%) | |
| In the last three months, how often did you feel heard and understood by your health care professionals? | Never | 49 (3.6%) |
| Sometimes | 113 (8.3%) | |
| Usually | 431 (31.6%) | |
| Always | 771 (56.5%) | |
| Care Coordination | ||
| In the last three months, how often did your health care professionals seem to know the important information about your medical history? | Never | 33 (2.4%) |
| Sometimes | 104 (7.6%) | |
| Usually | 446 (32.6%) | |
| Always | 787 (57.4%) | |
| In the last three months, did any of your health care professionals talk with you about the care or treatment you get from your other doctors or health care providers? | Yes, definitely | 522 (38.5%) |
| Yes, somewhat | 462 (34.1%) | |
| No | 372 (27.4%) | |
| In the last three months, did any of your health care professionals talk with you about all the medicines you are taking? | Yes, definitely | 940 (68.5%) |
| Yes, somewhat | 272 (19.8%) | |
| No | 153 (11.2%) | |
| I do not take any medicines | 7 (0.5%) | |
| In the last three months, did any of your health care professionals talk with you about how to get help with everyday activities? | Yes, definitely | 285 (29.1%) |
| Yes, somewhat | 280 (28.6%) | |
| No | 414 (42.3%) | |
| In the last three months, did you know who to call if you had questions or needed help with your care between visits? | Yes, definitely | 1038 (75.8%) |
| Yes, somewhat | 244 (17.8%) | |
| No | 88 (6.4%) | |
| In the last three months, when you contacted your health care professionals between visits for questions or help with your care, did you get the help you needed? | Yes, definitely | 761 (73.5%) |
| Yes, somewhat | 230 (22.2%) | |
| No | 44 (4.3%) | |
| Help for Symptoms | ||
| In the last three months, did you get as much help as you wanted for your pain? | Yes, definitely | 316 (45.1%) |
| Yes, somewhat | 283 (40.4%) | |
| No | 101 (14.4%) | |
| In the last three months, did you get as much help as you wanted for your breathing? | Yes, definitely | 208 (61.2%) |
| Yes, somewhat | 101 (29.7%) | |
| No | 31 (9.1%) | |
| In the last three months, did you get as much help as you wanted for your feelings of anxiety or sadness? | Yes, definitely | 179 (46.9%) |
| Yes, somewhat | 144 (37.7%) | |
| No | 59 (15.4%) | |
| Planning for Care | ||
| Did any of your health care professionals ever talk with you about what is important in your life? | Yes, definitely | 463 (34.0%) |
| Yes, somewhat | 427 (31.4%) | |
| No | 472 (34.7%) | |
| Did any of your health care professionals ever talk with you about what you should do during a health emergency? | Yes, definitely | 716 (52.3%) |
| Yes, somewhat | 285 (20.8%) | |
| No | 368 (26.9%) | |
| Did any of your health care professionals ever talk with you about what your health care options would be if you got sicker? | Yes, definitely | 411 (30.0%) |
| Yes, somewhat | 350 (25.6%) | |
| No | 607 (44.4%) | |
| Overall Rating of Care (single-item measure) | ||
| Using any number from 0 to 10, where 0 is the worst care possible and 10 is the best care possible, what number would you use to rate your care from your health care professionals? | 0 | 3 (0.2%) |
| 1 | 1 (0.1%) | |
| 2 | 2 (0.1%) | |
| 3 | 8 (0.6%) | |
| 4 | 10 (0.7%) | |
| 5 | 40 (2.9%) | |
| 6 | 37 (2.7%) | |
| 7 | 112 (8.2%) | |
| 8 | 301 (22.0%) | |
| 9 | 296 (21.6%) | |
| 10 | 560 (40.9%) |
Response distributions exclude respondents that did not respond to the given survey item or were ineligible to answer the item.