Evaluation of the Health and Economic Consequences of Kentucky's Section 1115 Demonstration Waiver

Kentucky HEALTH was an 1115 Medicaid waiver that was approved by the Centers for Medicare and Medicaid Services (CMS) in January 2018. In what was initially planned to be a 5-year demonstration, KY HEALTH aimed to modify the traditional Medicaid program to improve health behaviors, health outcomes, and socioeconomic outcomes in the waiver-eligible population through several innovations. In brief, these included introducing Community Engagement requirements (i.e. work requirements), monthly premiums, MyRewards accounts for dental and vision services, and annual recertification. If beneficiaries failed to complete these requirements, some penalties included suspension and 6-month lockouts from the Medicaid program.

The Commonwealth of Kentucky had chosen to implement this program in a randomized fashion, where 10% of the target population was randomly assigned to continue receiving traditional Medicaid while 90% would receive Kentucky HEALTH benefits and be subject to the requirements discussed above. Randomization was conducted by the state, through their separate contract with a non-profit research firm (National Opinion Research Center, NORC). The NORC also engaged in primary data collection to support the analysis of the project. The University of Pennsylvania team served as the non-partisan, independent evaluators of this randomized intervention conducted by the Commonwealth of Kentucky. The analysis would measure the impact of KY HEALTH compared to traditional Medicaid.

Due to ongoing legal challenges and a change in administrations, the implementation of Kentucky HEALTH was delayed and eventually cancelled. On March 27, 2019, the DC District Court Judge concluded that the approvals did not address how the requests would align with Medicaid's core objectives. At that point, data collection was ongoing although Kentucky HEALTH was delayed until further notice. Kentucky's 2019 gubernatorial election took place on November 5th and resulted in the election of a new governor. On December 16, 2019, a termination request was sent to CMS as a notification of the new administration's intention to cancel Kentucky HEALTH. On December 18, 2019, a clarification letter was sent to CMS to notify them that the termination did not apply to the SUD program and NEMT portion (among others) of the waiver. While the Penn team will continue conducting a separate evaluation of the ongoing SUD program, the randomized controlled trial and data collection described in this study will end with the terminated components of the waiver.

As a result of the legal challenges, the implementation of Kentucky HEALTH was delayed before being cancelled altogether. The study start and end dates above reflect the beginning of data collection and the termination letter that was sent to CMS, respectively.

A total of 9,396 surveys (KHES), 127 beneficiary semi-structured interviews, and 40 provider interviews were conducted.

Study Overview

• Status: Terminated
• Conditions: Health Insurance
• Intervention / Treatment: Other: Kentucky HEALTH

Detailed Description

General Background

With the approval of the Centers for Medicare and Medicaid Services (CMS) on January 12, 2018, the Commonwealth of Kentucky is implementing a 5-year § 1115 demonstration project entitled Kentucky HEALTH (for "Helping to Engage and Achieve Long-Term Health"). The project will modify the Medicaid program for eligible beneficiaries, which include the majority of able-bodied adults.

The Commonwealth's Medicaid demonstration waiver seeks to improve health behaviors, health outcomes, and socioeconomic outcomes in the waiver-eligible population through several innovations as described in the § 1115 CMS Medicaid waiver approval letter from January, 2018:

• The Kentucky HEALTH benefit plan for the expansion population will provide a comprehensive commercial insurance benefit package. Dental services, vision services, and over the counter medications will now be provided via the beneficiary's My Rewards Account.

• Kentucky HEALTH beneficiaries, with the exception of medically frail individuals, pregnant women, former foster youth, and children, will be required to make sliding scale flat rate monthly premium payments based on family income. As described in the application the plan is:

  • Under 25% FPL - $1.00 per month
  • 25-50% FPL - $4.00 per month
  • 51-100% FPL - $8.00 per month
  • 101-138% FPL - $15.00 per month
• Beneficiaries will be provided a 60-day grace period to make their required monthly premium payments - individuals who do not elect to pay a required premium payment within sixty calendar days from the due date will face a six-month non-payment penalty.

• The non-payment penalty varies based on whether the member has income above or below federal poverty:

  • Above 100% FPL - member will be disenrolled from Kentucky HEALTH; re-enrollment waiting period of six months
  • At or below 100% FPL - no impact on eligibility; State Plan copayments required for all services; $25 is deducted from the My Rewards Account; My Rewards Account is suspended - not able to use funds in account, not able to accrue funds in account
  • Medically Frail - no impact to eligibility; No copayments required; My Rewards Account is suspended - not able to use funds in account, not able to accrue funds in account
• Beneficiaries will have the opportunity to end their non-payment penalty period sooner than six months by paying past debt, paying the premium for the reinstatement month, and participating in a financial or health literacy course.

• Kentucky HEALTH will require that all able-bodied working age adult beneficiaries without dependents participate in the community engagement and employment initiative to maintain enrollment. For the first 3 months of enrollment, beneficiaries will not be required to report community engagement hours. Starting at 4 months of enrollment, non-exempt individuals will be required to participate in 80 hours of community engagement per calendar month.

  • Qualifying activities include employed work, job skills training, job search activities, education, community service, participation in substance use disorder treatment, caregiving services for non-dependents with a chronic health condition, and/or volunteering.
  • After one month of non-completion, beneficiaries will have a one-month opportunity to remedy their deficit or face a Community Engagement Suspension, or a suspension of all benefits. In order to avoid suspension, beneficiaries must be current on all hours required for the new month, and either (1) make up all deficit hours from the previous month before the end of the current month, or (2) complete an approved re-enrollment course. The re-enrollment course may be used once per 12-month period. Beneficiaries who do not cure the deficiency will be locked out from the Medicaid program. They may re-enroll at any time after completing 80 hours of community engagement within a 30-day period.
  • Beneficiaries who complete more than 80 hours in a given month may apply their excess hours to receive MyRewards Account credits and to remedy a deficit from the previous month, but they may not apply the excess hours forward to a future month.

• Implementation of penalties for failure to provide timely documentation for recertification of Medicaid eligibility.

  • Recertification is a yearly process for Medicaid recipients to recertify their household information and eligibility for benefits based on the certification period. Certification periods are 12 months from the first month of approved eligibility.
  • Eligible beneficiaries who do not provide documentation for recertification will face disenrollment from Medicaid, but are given a 90-day extension to recertify their eligibility for benefits. Beneficiaries who fail to do so within the 90-day period will become ineligible of receiving Medicaid benefits for a 6-month non-eligibility period. The same process applies to beneficiaries who fail to report changes in circumstance in the required reporting period.

• Establishment of the MyRewards account program for all adult beneficiaries who are currently making monthly premium payments and not subject to lockouts, as well as pregnant women.

  • Beneficiaries can earn MyRewards account credits through engagement in eligible wellness behaviors; they can have MyRewards credits deducted to penalize premium nonpayment and non-emergency use of the emergency department.
  • Beneficiaries may spend MyRewards account credits on vision care, dental services, and other eligible services.
• Elimination of retroactive eligibility for all Medicaid beneficiaries, with the exception of pregnant women and former foster care youth.
• Elimination of non-emergency ground transportation for beneficiaries subject to the alternative benefit plan, with the exception of 19- and 20-year old beneficiaries receiving EPSDT services.
• Offering premium assistance for beneficiaries seeking to enroll in employer-sponsored insurance (ESI). Enrolling in ESI with premium assistance will be optional for the first year that a beneficiary is in Kentucky HEALTH and concurrently eligible for ESI. In subsequent years of eligibility for ESI that meets state requirements, the beneficiary will be required to enroll in the ESI program in order to continue participating in Medicaid.
• Implementation of funded (up to $1,000) deductible accounts at the beginning of each benefit year. Deductible account dollars will be deducted as beneficiaries use health care throughout the year. At the conclusion of each program year, up to 50% of the remaining deductible account balance will be transferred to the member's MyRewards account, prorated to reflect the number of months with active enrollment (i.e., not suspended or locked out).

Collectively, these program modifications aim to incentivize healthy behaviors, appropriate health care utilization, and greater community and labor force participation, in order to lead to better health and socioeconomic outcomes. Figure 1 displays a driver diagram which clarifies these program objectives.

In addition to the above, the demonstration waiver will also include the following:

• Delivery system modifications to enhance access to mental health and substance use disorder treatments. Federal financial participation (FFP) will apply to substance use disorder treatment - including early intervention, in and outpatient therapy, residential treatment, managed withdrawal, medication assisted therapy, peer interventions, and crisis care support.
• Infrastructure changes to encourage prevention of chronic diseases

These elements are specifically designed to improve access to care and care options for individuals with substance use disorder and individuals with other chronic medical conditions, for which the Commonwealth suffers from a significant burden of disease (Dwyer-Lindgren 2016).

Overview of Analysis Plan

The proposed evaluation plan will leverage rich administrative and survey data sources along with robust quantitative and qualitative methods to assess the effects of the demonstration waiver on insurance coverage, health care utilization, health behaviors, socioeconomic outcomes, and a broad range of health outcomes (including substance use disorder). We will examine the consequences of the waiver both among individuals who will be moved from traditional Medicaid to the waiver at the start of the demonstration, as well as among low-income, able-bodied individuals as a whole (who may participate in the Medicaid program in the future). Importantly, we will track individuals from their start in this program for up to 5 years to examine how trajectories in terms of labor market and health outcomes evolve over time for waiver program participants compared to similar individuals on traditional Medicaid. It is crucial to emphasize the research team's role as neutral evaluators in this process.

In addition to assessing the impacts of the demonstration waiver as a whole, the evaluation will also seek to estimate the individual impacts of key waiver components, specifically community engagement requirements, premiums and lockouts, redetermination lockouts, and the My Rewards incentive program.

Specifically, we will achieve these objectives through conducting the following, the first of which is the subject of this clinicatrials.gov record:

• Secondary analysis of data collected as part of a randomized intervention initiated by the Commonwealth of Kentucky to evaluate effects of the waiver on coverage, utilization, health, and socioeconomic status among Medicaid beneficiaries eligible for the waiver at the time of implementation (April 1 2019). Randomization will be statewide and will apply to all beneficiaries in the waiver-eligible population at the time of implementation. Data sources for the RCT include administrative data maintained by the state; the Kentucky Health Experiment Survey (KHES), a prospective survey conducted for Kentucky HEALTH by the NORC; and biometric assessments among a subgroup of KHES respondents.
• Natural experiment and observational methods to estimate effects of program components-specifically premiums and lockouts, community engagement, redetermination lockouts, and increased coverage for substance use disorder services-on program participation and insurance coverage among those randomized to the waiver arm.
• Semi-structured qualitative interviews of key population groups (including current and former Medicaid beneficiaries, Medicaid administration staff, and providers).
• Natural experiment and observational methods to evaluate, for all low-income able-bodied adults in the Commonwealth, the consequences of the demonstration waiver (and specific waiver components) on (1) probability of Medicaid participation in the future and (2) insurance coverage, utilization, health, and socioeconomic status.

The first three elements will focus on the population of current adult Medicaid beneficiaries who will be exposed to the demonstration waiver at the time of initial implementation. The last element will focus on all low-income able-bodied individuals, who may both be current and future beneficiaries of the program.

The proposed evaluation will yield actionable information for policymakers both within the state and in other states where similar demonstration waivers are being proposed. The evaluation plan is innovative in its use of an RCT, which is the gold standard approach to achieving causal inference (Manning, Newhouse et al. 1987, Finkelstein, Taubman et al. 2012, Baicker, Taubman et al. 2013, Newhouse and Normand 2017). In addition, our mixed-methods will use natural experiments, which can also have high internal validity, alongside qualitative methods, which provide granular data on process and outcome measures that cannot be collected through surveys or administrative sources. Combining insights from quantitative and qualitative data will provide information on multiple dimensions of interest.

Multiple data sources will be used throughout the evaluation. The state of Kentucky will provide administrative data that will include Medicaid claims including prescriptions; vital statistics; wages & hours worked; education; state benefits enrollment data; records from Kentucky HEALTH related to premium payment, redetermination, etc. Additional administrative data will be made available from nationally curated databases including the State Inpatient Databases, the State Emergency Department Databases, the American Community Survey, and the Behavioral Risk Factor Surveillance Survey. The research team will add to these these administrative sources with the Kentucky Health Experiment Survey, qualitative interviews with key populations, and biometric data from Medicaid beneficiaries.

Built into the evaluation plan is the opportunity to identify, (pre-)specify, and conduct additional analyses in future years. For example, new analyses of interest may emerge from our Year 1 quantitative plan and qualitative interviews, or from changes in the demonstration waiver or its implementation. When new questions emerge, we will explicitly pre-specify any new prospective analyses we wish to conduct, and identify any retrospective analyses that were not pre-specified. In this context, the goal of the current document is to balance pre-specification (which minimizes prospects of data mining) and the opportunity to continually learn from the data in a policy relevant manner.

• Study Type: Observational
• Enrollment (Actual): 378828

Contacts and Locations

Study Locations

• United States
  • Kentucky
    • Frankfort, Kentucky, United States, 40621
      • Kentucky Cabinet for Health and Familty Services
  • Maryland
    • Bethesda, Maryland, United States, 20814
      • NORC at the University of Chicago

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years to 64 years (ADULT)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Probability Sample

Study Population

The study population includes all waiver-eligible adults who are currently enrolled in Medicaid in the Commonwealth of Kentucky (i.e. the majority of able-bodied adults that make up the expansion population).

Description

Inclusion Criteria:

• Waiver-eligible adults currently enrolled in Medicaid

Exclusion Criteria:

• Medically frail
• Pregnant
• Former foster youth (up to age 26)

Study Plan

How is the study designed?

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Kentucky HEALTH; This group will consist of 90% of eligible beneficiaries who will receive Kentucky HEALTH benefits throughout the 5-year demonstration waiver.	Other: Kentucky HEALTH; KY HEALTH is the modified Medicaid program including key components such as community engagement requirements, premiums and lockouts, redetermination lockouts, and the My Rewards incentive program.
Traditional Medicaid (control); This group will consist of 10% of eligible beneficiaries who will continue receiving traditional Medicaid benefits as in place before July 1, 2018 throughout the 5-year demonstration waiver.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Current Medicaid enrollment	Administrative data for randomized individuals. It will be measured by comparing changes in outcome between the treatment and control.	The data will be provided by Kentucky between July 1, 2018 to June 30, 2023. The evaluation team will receive data six months in arrears (i.e. will receive data on 1/1/19 that will run through 6/30/18).
Months uninsured in the past year	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Type of insurance coverage	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Number of ED visits	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Number of ED visits for ambulatory care sensitive conditions	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Hospitalizations	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Hospitalizations for ambulatory care sensitive conditions	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Number of dental care visits	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Substance use	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Blood pressure	Biometric assessment, subgroup of beneficiaries with baseline hypertension by self-report on KHES baseline survey. It will be measured by comparing changes in outcome between the treatment and control.	Biometrics were collected at the intended baseline (July 2018) and will be collected two years post-implementation (April 2021), and possibly 5 years post-implementation (April 2023).
Hemoglobin a1C	Biometric assessment, subgroup of beneficiaries with baseline diabetes by self-report on KHES baseline survey. It will be measured by comparing changes in outcome between the treatment and control.	Biometrics were collected at the intended baseline (July 2018) and will be collected two years post-implementation (April 2021), and possibly 5 years post-implementation (April 2023).
BMI	Biometric assessment, subgroup of beneficiaries with baseline diabetes or hypertension by self-report on KHES baseline survey. It will be measured by comparing changes in outcome between the treatment and control.	Biometrics were collected at the intended baseline (July 2018) and will be collected two years post-implementation (April 2021), and possibly 5 years post-implementation (April 2023).
Total cholesterol	Biometric assessment, subgroup of beneficiaries with baseline diabetes or hypertension by self-report on KHES baseline survey. It will be measured by comparing changes in outcome between the treatment and control.	Biometrics were collected at the intended baseline (July 2018) and will be collected two years post-implementation (April 2021), and possibly 5 years post-implementation (April 2023).
Family income	Administrative data for randomized individuals. It will be measured by comparing changes in outcome between the treatment and control.	The data will be provided by Kentucky between July 1, 2018 to June 30, 2023. The evaluation team will receive data six months in arrears (i.e. will receive data on 1/1/19 that will run through 6/30/18).
Currently employed	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Hours worked per week	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Banking	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Participation in education or job training (composite)	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Mortality	Administrative data for randomized individuals (vitals). It will be measured by comparing changes in outcome between the treatment and control.	The data will be provided by Kentucky between July 1, 2018 to June 30, 2023. The evaluation team will receive data six months in arrears (i.e. will receive data on 1/1/19 that will run through 6/30/18).
Days physical health not good in past month	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).
Days mental health not good in past month	KHES self-report. It will be measured by comparing changes in outcome between the treatment and control.	Survey data was collected at the intended baseline (July 2018). It will be collected in September 2019 and every subsequent April (2020, 2021, 2022, 2023).

Collaborators and Investigators

• Sponsor: University of Pennsylvania
• Collaborators: Columbia University; NORC at the University of Chicago
• Investigators: Principal Investigator: Kristen Underhill, JD, DPhil, Columbia University

Publications and helpful links

General Publications

• Linn KA, Underhill K, Dixon EL, Bair EF, Ferrell WJ, Montgomery ME, Volpp KG, Venkataramani AS. The design of a randomized controlled trial to evaluate multi-dimensional effects of a section 1115 Medicaid demonstration waiver with community engagement requirements. Contemp Clin Trials. 2020 Nov;98:106173. doi: 10.1016/j.cct.2020.106173. Epub 2020 Oct 7.

Study record dates

Study Major Dates

• Study Start (ACTUAL): April 1, 2018
• Primary Completion (ACTUAL): December 16, 2019
• Study Completion (ACTUAL): December 16, 2019

Study Registration Dates

• First Submitted: June 25, 2018
• First Submitted That Met QC Criteria: July 17, 2018
• First Posted (ACTUAL): July 26, 2018

Study Record Updates

• Last Update Posted (ACTUAL): February 5, 2020
• Last Update Submitted That Met QC Criteria: February 3, 2020
• Last Verified: February 1, 2020

More Information

Terms related to this study

• Other Study ID Numbers: 828923

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No