At the Intersection of Health, Health Care and Policy Cite this article as: Diana D. McDonnell and Carrie L. Graham Medicaid Beneficiaries In California Reported Less Positive Experiences When Assigned To A Managed Care Plan Health Affairs, 34, no.3 (2015):447-454 doi: 10.1377/hlthaff.2014.0528
The online version of this article, along with updated information and services, is available at: http://content.healthaffairs.org/content/34/3/447.full.html
For Reprints, Links & Permissions: http://healthaffairs.org/1340_reprints.php E-mail Alerts : http://content.healthaffairs.org/subscriptions/etoc.dtl To Subscribe: http://content.healthaffairs.org/subscriptions/online.shtml
Health Affairs is published monthly by Project HOPE at 7500 Old Georgetown Road, Suite 600, Bethesda, MD 20814-6133. Copyright © 2015 by Project HOPE - The People-to-People Health Foundation. As provided by United States copyright law (Title 17, U.S. Code), no part of Health Affairs may be reproduced, displayed, or transmitted in any form or by any means, electronic or mechanical, including photocopying or by information storage or retrieval systems, without prior written permission from the Publisher. All rights reserved.
Not for commercial use or unauthorized distribution Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
Health Plan Enrollment By Diana D. McDonnell and Carrie L. Graham 10.1377/hlthaff.2014.0528 HEALTH AFFAIRS 34, NO. 3 (2015): 447–454 ©2015 Project HOPE— The People-to-People Health Foundation, Inc.
doi:
AG I NG
&
H E A LT H
Medicaid Beneficiaries In California Reported Less Positive Experiences When Assigned To A Managed Care Plan
Diana D. McDonnell ([email protected]) is an epidemiologist at the Center for Family and Community Health, School of Public Health, University of California, Berkeley.
In 2011 California began transitioning approximately 340,000 seniors and people with disabilities from Medicaid fee-for-service (FFS) to Medicaid managed care plans. When beneficiaries did not actively choose a managed care plan, the state assigned them to one using an algorithm based on their previous FFS primary and specialty care use. When no clear link could be established, beneficiaries were assigned by default to a managed care plan based on weighted randomization. In this article we report the results of a telephone survey of 1,521 seniors and people with disabilities enrolled in Medi-Cal (California Medicaid) and who were recently transitioned to a managed care plan. We found that 48 percent chose their own plan, 11 percent were assigned to a plan by algorithm, and 41 percent were assigned to a plan by default. People in the latter two categories reported being similarly less positive about their experiences compared to beneficiaries who actively chose a plan. Many states in addition to California are implementing mandatory transitions of Medicaid-only beneficiaries to managed care plans. Our results highlight the importance of encouraging beneficiaries to actively choose their health plan; when beneficiaries do not choose, states should employ robust intelligent assignment algorithms. ABSTRACT
M
edi-Cal, California’s Medicaid program, plays an important role in the lives of low-income seniors and people with disabilities. It acts as a medical safety net for nearly one in five California seniors, providing them with health care, medication, medical equipment, and long-term services and supports.1 Many beneficiaries have complex medical needs: Nearly half have three or more chronic conditions.2 In 1994 California began a mandatory transition of most Medi-Cal-only beneficiaries (those not enrolled in both Medicaid and Medicare) from fee-for-service (FFS) to Medi-Cal managed
Carrie L. Graham is assistant director of research at Health Research for Action, School of Public Health, University of California, Berkeley.
care plans.1 In addition to slowing long-term health care expenditures, this transition was intended to create more accountable and coordinated systems of care, strengthen the health care safety net, reward quality, and improve health outcomes.3 Managed care is now the dominant delivery system for most states’ Medicaid programs.4 Two-thirds of all Medicaid enrollees nationwide now receive most or all of their benefits through managed care. In addition to California, many other states, including Texas and New York, have begun mandatory transitions of their remaining FFS Medicaid beneficiaries to managed care. At the start of California’s transition in 1994, M a r ch 20 1 5
3 4 :3
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
Health A ffairs
447
Health Plan Enrollment Medi-Cal seniors and people with disabilities had been permitted to remain in FFS because of concerns about access to and quality of care for this medically complex population.5 In June 2011 the California Department of Health Care Services (DHCS) began mandatory transitions of approximately 340,000 seniors and people with disabilities to managed care as mandated by the state’s Section 1115 Bridge to Reform waiver and to take advantage of managed care’s potential for more coordinated and accountable health care delivery.3 During the first year of this transition, about 20,000 seniors and people with disabilities were exempted because of emergency disenrollment, medical exemption, or other reasons.6 However, by May 2012 about 240,000 of these beneficiaries had completed the process.6 Sixteen counties in California each offered beneficiaries a choice of at least two Medi-Cal managed care plans. Beneficiaries were notified by letter ninety, sixty, and thirty days prior to the transition. In addition to information about their county’s plans, the letters informed beneficiaries that they could either choose or be assigned to a plan. Beneficiaries who had not enrolled as their deadline approached received a telephone call to encourage them to select a plan. The number of beneficiaries who actively chose a plan was one of the metrics the state used to evaluate the transition because active choice has been associated with improved access to and satisfaction with care. For example, Claudia Schur and Marc Berk7 found that Medicaid enrollees who had a choice of plans obtained services more easily, rated plan quality higher, and reported fewer problems with plan rules compared to enrollees without a choice. Karen Davis and colleagues8 found that plan satisfaction was twice as high among managed care enrollees who had a choice of plans compared to those who did not. Beneficiaries who choose their own plans also are more likely to stay enrolled than those who do not choose.9,10 Despite the benefits of choosing their own plans, many people do not do so. A common way to address this lack of choice is to use “intelligent assignment” algorithms that are designed to select plans for nonchoosing beneficiaries based on their previous health care use, the availability and characteristics of available providers, and other data. As of July 2014 twelve states, including California, had received approval from the Centers for Medicare and Medicaid Services (CMS) to conduct demonstration projects to transition beneficiaries enrolled in both Medicare and Medicaid (dual eligibles) into managed care plans.11 CMS requires states to develop and use intelligent assignment algorithms to transi448
H ea lt h A f fai r s
Ma rc h 2 015
34 : 3
tion nonchoosing beneficiaries to a managed care plan.12 At the time of our study and to this present day, however, no specific processes or requirements for these algorithms have been set forth, and states have been on their own to develop and evaluate their own techniques. We could find nothing in the peer-reviewed literature evaluating or describing assignment algorithms in California or other states. Even when discussing the benefits of plan choice, many studies identify plan characteristics that are important to consumers’ enrollment decisions13 but do not evaluate beneficiaries’ subsequent experiences based on plan allocation. The current study was conducted among a representative sample of Medi-Cal seniors and people with disabilities who had transitioned from FFS to managed care from June 2011 to March 2012. We sought to determine how their experiences with the transition differed based on their choosing status. To do this, we asked about their experiences with the transition process, access to care in the managed care plan, and satisfaction with care. We then compared those who chose a plan (choosers) to those who were assigned a plan based on FFS utilization data (assigned) and to those who were assigned by default to a plan because of incomplete or inconclusive FFS data (defaulters). Our hypothesis was that using FFS data to assign beneficiaries to plans would replicate the process of active choice, and, thus, the assigned beneficiaries would report experiences that were similar to those of active choosers and significantly different from beneficiaries classified as defaulters. However, this proved not to be the case: Those who were assigned to a plan reported outcomes similar to those who defaulted into a plan, with beneficiaries in both groups being far less satisfied than those who actively chose a plan.
Study Data And Methods A telephone survey was conducted in English and Spanish with 1,521 recently transitioned seniors and people with disabilities or their health care proxies. The main study is described elsewhere.14 Briefly, the study assessed beneficiaries’ self-reported experiences with notification, enrollment, and care during their first months in managed care. The survey was informed by instruments used in other studies that examined similar managed care transitions5,15–19 and was reviewed by an advisory group of health care providers, Medi-Cal health plan administrators, and patient advocates. It was administered between October and December 2012 using computer-assisted tele-
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
phone interviewing. Interviews lasted approximately twenty minutes. Participants were selected from a California DHCS list of seniors and people with disabilities age eighteen or older who completed mandatory transitions to managed care from June 2011 to March 2012. Eligible beneficiaries were not enrolled in Medicare, meaning that seniors in our study (about 11 percent of the sample) were not eligible for Medicare, typically because neither they nor their spouses had worked the approximately ten years of Medicare-covered employment required to be eligible. People in this situation often received income that was unreported or immigrated to the United States later in life. Participants also had to have been in FFS Medi-Cal for at least six months prior to their managed care transition because only brief experience with FFS Medi-Cal would decrease their ability to compare FFS and managed care (about 11 percent of otherwise eligible people were excluded for this reason). Potential respondents also needed to have a valid telephone number and address on file with the California DHCS because letters were sent to them prior to the study. Beneficiaries who lived in group homes or other residential care settings were excluded because they were hypothesized to have different health information and care experiences. Of the 172,334 eligible beneficiaries, the California DHCS randomly selected 29,493 based on a conservative estimate of the response rate and desired final sample of 1,515. Stratified sampling was used to ensure adequate respondents in each plan allocation group, described below. Calls were made to 5,101 potential respondents. Approximately one-half (52 percent) were ineligible, mostly because they were no longer on Medi-Cal, did not speak English or Spanish, or had nonworking telephone numbers. Of the remaining 2,466, 38 percent declined to participate, yielding a final sample of 1,521 for an American Association for Public Opinion Research response rate20 of 62 percent. Plan Allocation In the notification letters, beneficiaries were asked to choose between at least two Medi-Cal managed care plans offered in their county. For this analysis, those who chose by the specified deadline (first day of their birth month) were classified as “choosers.” For those who did not choose, the California DHCS sought to assign them to a health plan with which they had some relationship through an affiliated provider (intelligent assignment). Using Medi-Cal FFS data from providers rendering both primary and specialty care during the previous twelve months, the state attempted to identify one “top provider” for each beneficiary based on number of visits and reimbursed ex-
penses, with priority given to the provider who saw the beneficiary most. If the top provider belonged to a managed care plan in the beneficiary’s county, the beneficiary was assigned to that plan. For this analysis, beneficiaries in this group were classified as “assigned.” When the identified top provider was affiliated with more than one plan, was not affiliated with any plan, or could not be identified, that beneficiary was assigned based on a “default algorithm”—a weighted randomization with assignment probability proportional to a plan’s quality rating, based on the Healthcare Effectiveness Data and Information Set (HEDIS) and safetynet measures.21 For this analysis, those in this group were classified as “defaulters.” Analysis Our aim was to evaluate whether assigning nonchoosing beneficiaries to a managed care plan based on their top provider resulted in experiences with transitioning to and receiving services in managed care comparable to the experiences of active choosers. The assigned were, therefore, compared to both choosers (the benchmark) and defaulters (how they would have been assigned absent the intervention). Outcome variables (all self-reported) are listed in online Appendix Exhibit A1.22 Bivariate comparisons were tested for statistical significance using chi-square testing for categorical and analysis of variance (ANOVA) for continuous variables. Logistic regression models were used to assess independent effects of plan allocation on transition and care outcomes. Models controlled for demographics, health status, health care use, health conditions, and Medi-Cal enrollment information. The former three types of variables were self-reported, and the latter two were obtained from Medi-Cal administrative data. Comparisons were considered statistically significant at the p < 0:05 level; because of many comparisons, significance values of p < 0:01 were also noted. All analyses were performed using data weighted based on sampling probability and nonresponse, in SPSS version 20. Limitations Several limitations warrant mention. Although information about plan assignment was obtained from administrative data, all outcomes were self-reported, meaning that we are specifically assessing people's perceptions about the transition. The study also included only beneficiaries with valid contact information who could communicate in English or Spanish, and results can, therefore, be generalized only to those groups. Regarding plan allocation, when a beneficiary’s top provider was not clearly linked to a single county plan, he or she was defaulted. There was no attempt to assess the affiliations March 2015
34:3
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
H ea lt h A f fai r s
44 9
Health Plan Enrollment of additional providers. Furthermore, top providers were selected from primary or specialty care. Ancillary providers such as laboratories and pharmacies were not included in the algorithm. This study also only included people in California. The generalizability of our findings could be limited to the extent that Medicaid recipients in California differ from those in other states or by the existence of state-specific processes for enrollment (such as limited marketing efforts by managed care plans). Although this study was a first step in evaluating the efficacy of efforts to intelligently assign beneficiaries to health plans, it was not exhaustive. For example, although the multivariate models controlled for demographics and health status, there could be other differences between plan allocation groups that we did not measure and that could help explain the more positive experiences reported by choosers, perhaps related to personality type, social support, or patient activation.23 Similarly, this study focused on beneficiaries’ immediate experiences with enrollment and care; we did not assess long-term health outcomes.
Study Results While more than half of the beneficiaries in the study (52 percent) did not choose a plan, only 11 percent were assigned based on FFS data. The rest (41 percent) were assigned based on the default algorithm (Exhibit 1). Comparing Choosers And Assigned The chooser and assigned groups had similar demographic profiles, but assigned beneficiaries generally reported worse health (Exhibit 2). For example, assigned beneficiaries were more likely to report difficulty getting to places outside of walking distance (63 percent assigned, 51 percent choosers), having five or more primary care physician visits during the previous six months
Exhibit 1 Distribution Of Study Sample Of Medi-Cal Seniors And People With Disabilities Who Transitioned From Fee-For-Service To Managed Care, June 2011–March 2012 Unweighted
Weighted
Stratum Chose
Description Chose plan on their own
Percent 34
Number 519
Percent 48
Number 581
Assigned Default
Top provider in one plan Top provider not in any plan, in more than one plan, or not identified
16 50
235 767
11 41
128 492
SOURCE Authors’ analyses. NOTES Plan allocation strata are defined in the text. N ¼ 1; 201 (weighted data).
450
H e a lt h A f fai r s
March 2015
34:3
(32 percent, 24 percent), and having three or more emergency department (ED) visits during the previous six months (18 percent, 10 percent). Significantly more of those who were assigned (94 percent) compared to those who chose (88 percent) had Medi-Cal records indicating claims for chronic conditions during the previous year. When reporting experiences with Medi-Cal managed care enrollment, choosers were significantly more likely than the assigned to have enrolled early in the transition process. Before they transitioned, choosers were more likely (22 percent) than assigned (14 percent) to have been in Medi-Cal for less than twenty-four months. There were no significant differences between choosers and assigned by plan category (local initiative or commercial plan), plan name, or county of residence. Comparing experiences with the transition, choosers generally reported more positive outcomes than those who were assigned. For example, the assigned (13 percent) were significantly more likely than choosers (5 percent) to not remember being notified about the transition, find the information they received about the transition not at all useful (32 percent, 18 percent), and not understand that they could choose between plans (26 percent, 15 percent). Despite these differences, there was much that did not differ between the assigned and choosers. For example, there were no differences in the proportion reporting that it was very difficult to get information about their plan, receiving the provider they chose at enrollment, or remembering a call from their new plan. When reporting their experiences with MediCal managed care, the assigned were less likely to know how to navigate their new plan compared to those who chose. For example, they were significantly more likely than choosers to not know how to find a provider (44 percent assigned, 28 percent choosers), make an appointment for primary care (23 percent, 13 percent), get medical tests (33 percent, 23 percent), or get phone advice (38 percent, 29 percent). More of the assigned than choosers reported that it was difficult to get appointments with primary care providers (26 percent, 15 percent) and specialists (27 percent, 16 percent). The assigned also more often reported that the quality of their care was worse following the transition (29 percent, 17 percent), that they received less help finding a doctor and getting services (32 percent, 18 percent), and that they had had difficulty being seen by a doctor because the office lacked disability access (14 percent, 8 percent). Overall, significantly more assigned (44 per-
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
Exhibit 2 Bivariate Analysis: Medi-Cal Seniors And People With Disabilities Who Transitioned From Fee-For-Service To Managed Care, By Plan Allocation, June 2011–March 2012 p values
Plan allocation stratum
Descriptive characteristics Health status Outside walking distance difficult Health care use in previous six months 5+ PCP visits 3+ ED visits Health conditions Any chronic condition claim Medi-Cal information Period enrolled Jun–Aug 2011 Sep–Dec 2011 Jan–Mar 2012 Medi-Cal eligible 136 months Enrollment to survey, mean days Outcomes Experience with notification and enrollment Do not recall notification Information not at all useful Did not understand that they could choose Did not try to find plan information Did not enroll in original plan Experience with care Did not know how to: Find doctor Obtain lab tests Schedule PCP appointment Get phone advice Switch plans Benefits were worse than before Getting PCP appointment was more difficult Getting specialist appointment was more difficult Quality of care was worse Received less help getting services Ever had doctor lack access for disabled
Chose (n=581)
Assigned (n=128)
Default (n=492)
Assigned vs. chose
Assigned vs. default
51%
63%
57%
0.021
0.327
24 10
32 18
30 17
0.042 0.031
0.179 0.281
88
94
91
0.046
0.190
30 43 27
6 52 43
14 46 40
The online version of this article, along with updated information and services, is available at: http://content.healthaffairs.org/content/34/3/447.full.html
For Reprints, Links & Permissions: http://healthaffairs.org/1340_reprints.php E-mail Alerts : http://content.healthaffairs.org/subscriptions/etoc.dtl To Subscribe: http://content.healthaffairs.org/subscriptions/online.shtml
Health Affairs is published monthly by Project HOPE at 7500 Old Georgetown Road, Suite 600, Bethesda, MD 20814-6133. Copyright © 2015 by Project HOPE - The People-to-People Health Foundation. As provided by United States copyright law (Title 17, U.S. Code), no part of Health Affairs may be reproduced, displayed, or transmitted in any form or by any means, electronic or mechanical, including photocopying or by information storage or retrieval systems, without prior written permission from the Publisher. All rights reserved.
Not for commercial use or unauthorized distribution Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
Health Plan Enrollment By Diana D. McDonnell and Carrie L. Graham 10.1377/hlthaff.2014.0528 HEALTH AFFAIRS 34, NO. 3 (2015): 447–454 ©2015 Project HOPE— The People-to-People Health Foundation, Inc.
doi:
AG I NG
&
H E A LT H
Medicaid Beneficiaries In California Reported Less Positive Experiences When Assigned To A Managed Care Plan
Diana D. McDonnell ([email protected]) is an epidemiologist at the Center for Family and Community Health, School of Public Health, University of California, Berkeley.
In 2011 California began transitioning approximately 340,000 seniors and people with disabilities from Medicaid fee-for-service (FFS) to Medicaid managed care plans. When beneficiaries did not actively choose a managed care plan, the state assigned them to one using an algorithm based on their previous FFS primary and specialty care use. When no clear link could be established, beneficiaries were assigned by default to a managed care plan based on weighted randomization. In this article we report the results of a telephone survey of 1,521 seniors and people with disabilities enrolled in Medi-Cal (California Medicaid) and who were recently transitioned to a managed care plan. We found that 48 percent chose their own plan, 11 percent were assigned to a plan by algorithm, and 41 percent were assigned to a plan by default. People in the latter two categories reported being similarly less positive about their experiences compared to beneficiaries who actively chose a plan. Many states in addition to California are implementing mandatory transitions of Medicaid-only beneficiaries to managed care plans. Our results highlight the importance of encouraging beneficiaries to actively choose their health plan; when beneficiaries do not choose, states should employ robust intelligent assignment algorithms. ABSTRACT
M
edi-Cal, California’s Medicaid program, plays an important role in the lives of low-income seniors and people with disabilities. It acts as a medical safety net for nearly one in five California seniors, providing them with health care, medication, medical equipment, and long-term services and supports.1 Many beneficiaries have complex medical needs: Nearly half have three or more chronic conditions.2 In 1994 California began a mandatory transition of most Medi-Cal-only beneficiaries (those not enrolled in both Medicaid and Medicare) from fee-for-service (FFS) to Medi-Cal managed
Carrie L. Graham is assistant director of research at Health Research for Action, School of Public Health, University of California, Berkeley.
care plans.1 In addition to slowing long-term health care expenditures, this transition was intended to create more accountable and coordinated systems of care, strengthen the health care safety net, reward quality, and improve health outcomes.3 Managed care is now the dominant delivery system for most states’ Medicaid programs.4 Two-thirds of all Medicaid enrollees nationwide now receive most or all of their benefits through managed care. In addition to California, many other states, including Texas and New York, have begun mandatory transitions of their remaining FFS Medicaid beneficiaries to managed care. At the start of California’s transition in 1994, M a r ch 20 1 5
3 4 :3
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
Health A ffairs
447
Health Plan Enrollment Medi-Cal seniors and people with disabilities had been permitted to remain in FFS because of concerns about access to and quality of care for this medically complex population.5 In June 2011 the California Department of Health Care Services (DHCS) began mandatory transitions of approximately 340,000 seniors and people with disabilities to managed care as mandated by the state’s Section 1115 Bridge to Reform waiver and to take advantage of managed care’s potential for more coordinated and accountable health care delivery.3 During the first year of this transition, about 20,000 seniors and people with disabilities were exempted because of emergency disenrollment, medical exemption, or other reasons.6 However, by May 2012 about 240,000 of these beneficiaries had completed the process.6 Sixteen counties in California each offered beneficiaries a choice of at least two Medi-Cal managed care plans. Beneficiaries were notified by letter ninety, sixty, and thirty days prior to the transition. In addition to information about their county’s plans, the letters informed beneficiaries that they could either choose or be assigned to a plan. Beneficiaries who had not enrolled as their deadline approached received a telephone call to encourage them to select a plan. The number of beneficiaries who actively chose a plan was one of the metrics the state used to evaluate the transition because active choice has been associated with improved access to and satisfaction with care. For example, Claudia Schur and Marc Berk7 found that Medicaid enrollees who had a choice of plans obtained services more easily, rated plan quality higher, and reported fewer problems with plan rules compared to enrollees without a choice. Karen Davis and colleagues8 found that plan satisfaction was twice as high among managed care enrollees who had a choice of plans compared to those who did not. Beneficiaries who choose their own plans also are more likely to stay enrolled than those who do not choose.9,10 Despite the benefits of choosing their own plans, many people do not do so. A common way to address this lack of choice is to use “intelligent assignment” algorithms that are designed to select plans for nonchoosing beneficiaries based on their previous health care use, the availability and characteristics of available providers, and other data. As of July 2014 twelve states, including California, had received approval from the Centers for Medicare and Medicaid Services (CMS) to conduct demonstration projects to transition beneficiaries enrolled in both Medicare and Medicaid (dual eligibles) into managed care plans.11 CMS requires states to develop and use intelligent assignment algorithms to transi448
H ea lt h A f fai r s
Ma rc h 2 015
34 : 3
tion nonchoosing beneficiaries to a managed care plan.12 At the time of our study and to this present day, however, no specific processes or requirements for these algorithms have been set forth, and states have been on their own to develop and evaluate their own techniques. We could find nothing in the peer-reviewed literature evaluating or describing assignment algorithms in California or other states. Even when discussing the benefits of plan choice, many studies identify plan characteristics that are important to consumers’ enrollment decisions13 but do not evaluate beneficiaries’ subsequent experiences based on plan allocation. The current study was conducted among a representative sample of Medi-Cal seniors and people with disabilities who had transitioned from FFS to managed care from June 2011 to March 2012. We sought to determine how their experiences with the transition differed based on their choosing status. To do this, we asked about their experiences with the transition process, access to care in the managed care plan, and satisfaction with care. We then compared those who chose a plan (choosers) to those who were assigned a plan based on FFS utilization data (assigned) and to those who were assigned by default to a plan because of incomplete or inconclusive FFS data (defaulters). Our hypothesis was that using FFS data to assign beneficiaries to plans would replicate the process of active choice, and, thus, the assigned beneficiaries would report experiences that were similar to those of active choosers and significantly different from beneficiaries classified as defaulters. However, this proved not to be the case: Those who were assigned to a plan reported outcomes similar to those who defaulted into a plan, with beneficiaries in both groups being far less satisfied than those who actively chose a plan.
Study Data And Methods A telephone survey was conducted in English and Spanish with 1,521 recently transitioned seniors and people with disabilities or their health care proxies. The main study is described elsewhere.14 Briefly, the study assessed beneficiaries’ self-reported experiences with notification, enrollment, and care during their first months in managed care. The survey was informed by instruments used in other studies that examined similar managed care transitions5,15–19 and was reviewed by an advisory group of health care providers, Medi-Cal health plan administrators, and patient advocates. It was administered between October and December 2012 using computer-assisted tele-
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
phone interviewing. Interviews lasted approximately twenty minutes. Participants were selected from a California DHCS list of seniors and people with disabilities age eighteen or older who completed mandatory transitions to managed care from June 2011 to March 2012. Eligible beneficiaries were not enrolled in Medicare, meaning that seniors in our study (about 11 percent of the sample) were not eligible for Medicare, typically because neither they nor their spouses had worked the approximately ten years of Medicare-covered employment required to be eligible. People in this situation often received income that was unreported or immigrated to the United States later in life. Participants also had to have been in FFS Medi-Cal for at least six months prior to their managed care transition because only brief experience with FFS Medi-Cal would decrease their ability to compare FFS and managed care (about 11 percent of otherwise eligible people were excluded for this reason). Potential respondents also needed to have a valid telephone number and address on file with the California DHCS because letters were sent to them prior to the study. Beneficiaries who lived in group homes or other residential care settings were excluded because they were hypothesized to have different health information and care experiences. Of the 172,334 eligible beneficiaries, the California DHCS randomly selected 29,493 based on a conservative estimate of the response rate and desired final sample of 1,515. Stratified sampling was used to ensure adequate respondents in each plan allocation group, described below. Calls were made to 5,101 potential respondents. Approximately one-half (52 percent) were ineligible, mostly because they were no longer on Medi-Cal, did not speak English or Spanish, or had nonworking telephone numbers. Of the remaining 2,466, 38 percent declined to participate, yielding a final sample of 1,521 for an American Association for Public Opinion Research response rate20 of 62 percent. Plan Allocation In the notification letters, beneficiaries were asked to choose between at least two Medi-Cal managed care plans offered in their county. For this analysis, those who chose by the specified deadline (first day of their birth month) were classified as “choosers.” For those who did not choose, the California DHCS sought to assign them to a health plan with which they had some relationship through an affiliated provider (intelligent assignment). Using Medi-Cal FFS data from providers rendering both primary and specialty care during the previous twelve months, the state attempted to identify one “top provider” for each beneficiary based on number of visits and reimbursed ex-
penses, with priority given to the provider who saw the beneficiary most. If the top provider belonged to a managed care plan in the beneficiary’s county, the beneficiary was assigned to that plan. For this analysis, beneficiaries in this group were classified as “assigned.” When the identified top provider was affiliated with more than one plan, was not affiliated with any plan, or could not be identified, that beneficiary was assigned based on a “default algorithm”—a weighted randomization with assignment probability proportional to a plan’s quality rating, based on the Healthcare Effectiveness Data and Information Set (HEDIS) and safetynet measures.21 For this analysis, those in this group were classified as “defaulters.” Analysis Our aim was to evaluate whether assigning nonchoosing beneficiaries to a managed care plan based on their top provider resulted in experiences with transitioning to and receiving services in managed care comparable to the experiences of active choosers. The assigned were, therefore, compared to both choosers (the benchmark) and defaulters (how they would have been assigned absent the intervention). Outcome variables (all self-reported) are listed in online Appendix Exhibit A1.22 Bivariate comparisons were tested for statistical significance using chi-square testing for categorical and analysis of variance (ANOVA) for continuous variables. Logistic regression models were used to assess independent effects of plan allocation on transition and care outcomes. Models controlled for demographics, health status, health care use, health conditions, and Medi-Cal enrollment information. The former three types of variables were self-reported, and the latter two were obtained from Medi-Cal administrative data. Comparisons were considered statistically significant at the p < 0:05 level; because of many comparisons, significance values of p < 0:01 were also noted. All analyses were performed using data weighted based on sampling probability and nonresponse, in SPSS version 20. Limitations Several limitations warrant mention. Although information about plan assignment was obtained from administrative data, all outcomes were self-reported, meaning that we are specifically assessing people's perceptions about the transition. The study also included only beneficiaries with valid contact information who could communicate in English or Spanish, and results can, therefore, be generalized only to those groups. Regarding plan allocation, when a beneficiary’s top provider was not clearly linked to a single county plan, he or she was defaulted. There was no attempt to assess the affiliations March 2015
34:3
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
H ea lt h A f fai r s
44 9
Health Plan Enrollment of additional providers. Furthermore, top providers were selected from primary or specialty care. Ancillary providers such as laboratories and pharmacies were not included in the algorithm. This study also only included people in California. The generalizability of our findings could be limited to the extent that Medicaid recipients in California differ from those in other states or by the existence of state-specific processes for enrollment (such as limited marketing efforts by managed care plans). Although this study was a first step in evaluating the efficacy of efforts to intelligently assign beneficiaries to health plans, it was not exhaustive. For example, although the multivariate models controlled for demographics and health status, there could be other differences between plan allocation groups that we did not measure and that could help explain the more positive experiences reported by choosers, perhaps related to personality type, social support, or patient activation.23 Similarly, this study focused on beneficiaries’ immediate experiences with enrollment and care; we did not assess long-term health outcomes.
Study Results While more than half of the beneficiaries in the study (52 percent) did not choose a plan, only 11 percent were assigned based on FFS data. The rest (41 percent) were assigned based on the default algorithm (Exhibit 1). Comparing Choosers And Assigned The chooser and assigned groups had similar demographic profiles, but assigned beneficiaries generally reported worse health (Exhibit 2). For example, assigned beneficiaries were more likely to report difficulty getting to places outside of walking distance (63 percent assigned, 51 percent choosers), having five or more primary care physician visits during the previous six months
Exhibit 1 Distribution Of Study Sample Of Medi-Cal Seniors And People With Disabilities Who Transitioned From Fee-For-Service To Managed Care, June 2011–March 2012 Unweighted
Weighted
Stratum Chose
Description Chose plan on their own
Percent 34
Number 519
Percent 48
Number 581
Assigned Default
Top provider in one plan Top provider not in any plan, in more than one plan, or not identified
16 50
235 767
11 41
128 492
SOURCE Authors’ analyses. NOTES Plan allocation strata are defined in the text. N ¼ 1; 201 (weighted data).
450
H e a lt h A f fai r s
March 2015
34:3
(32 percent, 24 percent), and having three or more emergency department (ED) visits during the previous six months (18 percent, 10 percent). Significantly more of those who were assigned (94 percent) compared to those who chose (88 percent) had Medi-Cal records indicating claims for chronic conditions during the previous year. When reporting experiences with Medi-Cal managed care enrollment, choosers were significantly more likely than the assigned to have enrolled early in the transition process. Before they transitioned, choosers were more likely (22 percent) than assigned (14 percent) to have been in Medi-Cal for less than twenty-four months. There were no significant differences between choosers and assigned by plan category (local initiative or commercial plan), plan name, or county of residence. Comparing experiences with the transition, choosers generally reported more positive outcomes than those who were assigned. For example, the assigned (13 percent) were significantly more likely than choosers (5 percent) to not remember being notified about the transition, find the information they received about the transition not at all useful (32 percent, 18 percent), and not understand that they could choose between plans (26 percent, 15 percent). Despite these differences, there was much that did not differ between the assigned and choosers. For example, there were no differences in the proportion reporting that it was very difficult to get information about their plan, receiving the provider they chose at enrollment, or remembering a call from their new plan. When reporting their experiences with MediCal managed care, the assigned were less likely to know how to navigate their new plan compared to those who chose. For example, they were significantly more likely than choosers to not know how to find a provider (44 percent assigned, 28 percent choosers), make an appointment for primary care (23 percent, 13 percent), get medical tests (33 percent, 23 percent), or get phone advice (38 percent, 29 percent). More of the assigned than choosers reported that it was difficult to get appointments with primary care providers (26 percent, 15 percent) and specialists (27 percent, 16 percent). The assigned also more often reported that the quality of their care was worse following the transition (29 percent, 17 percent), that they received less help finding a doctor and getting services (32 percent, 18 percent), and that they had had difficulty being seen by a doctor because the office lacked disability access (14 percent, 8 percent). Overall, significantly more assigned (44 per-
Downloaded from content.healthaffairs.org by Health Affairs on June 21, 2015 at Univ of Connecticut Health Center/L.M. Stowe Library
Exhibit 2 Bivariate Analysis: Medi-Cal Seniors And People With Disabilities Who Transitioned From Fee-For-Service To Managed Care, By Plan Allocation, June 2011–March 2012 p values
Plan allocation stratum
Descriptive characteristics Health status Outside walking distance difficult Health care use in previous six months 5+ PCP visits 3+ ED visits Health conditions Any chronic condition claim Medi-Cal information Period enrolled Jun–Aug 2011 Sep–Dec 2011 Jan–Mar 2012 Medi-Cal eligible 136 months Enrollment to survey, mean days Outcomes Experience with notification and enrollment Do not recall notification Information not at all useful Did not understand that they could choose Did not try to find plan information Did not enroll in original plan Experience with care Did not know how to: Find doctor Obtain lab tests Schedule PCP appointment Get phone advice Switch plans Benefits were worse than before Getting PCP appointment was more difficult Getting specialist appointment was more difficult Quality of care was worse Received less help getting services Ever had doctor lack access for disabled
Chose (n=581)
Assigned (n=128)
Default (n=492)
Assigned vs. chose
Assigned vs. default
51%
63%
57%
0.021
0.327
24 10
32 18
30 17
0.042 0.031
0.179 0.281
88
94
91
0.046
0.190
30 43 27
6 52 43
14 46 40
