TEXAS - cdn.ymaws.com

i

PREPARED FOR

Methodist Healthcare Ministries

of South Texas, Inc.

by UT School of Public Health

Charles Begley, Ph.D., Lee Revere, Ph.D., Ellerie Weber, Ph.D., Alissa Ratanatawan, Ph.D. Candidate, Youngran Kim, Ph.D. Candidate, Hsiao Ling Phuar, Ph.D. Candidate

August 2016

TEXAS MEDICAID PERFORMANCE STUDY

2 0 1 6 F I N A L R E P O R T

ACKNOWLED GEMENT

Methodist Healthcare Ministries of South Texas, Inc. gratefully

acknowledges The University of Texas Health Science Center of

Houston School of Public Health and each of the researchers who

contributed their time, expertise and skill in researching and producing

the data and content featured in this report, most especially

Charles Begley, Ph.D., Lee Revere, Ph.D., Ellerie Weber, Ph.D.,

Alissa Ratanatawan, Ph.D. Candidate, Youngran Kim, Ph.D. Candidate

and Hsiao Ling Phuar, Ph.D. Candidate.

Methodist Healthcare Ministries also gratefully acknowledges the

Texas Association of Community Health Plans for their support of this

effort and the technical assistance provided to the researchers.

1

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–7

Well-Child Visits in the First 15 Months of Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

HEDIS Measures of Access and Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Related Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6–7

Well Child Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6–7

Cesarean Section Deliveries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Asthma Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8–11

Truven Health Analytic MarketScan Commercial Claims and Encounters . . . . . . . . . . . . . . . . . . . 8–9

Texas Medicaid Claims and Encounters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Access and Quality Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Well Child Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Prenatal Visit and Cesarean Section Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Asthma Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10–11

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12–19

Well-Child Visits in the First 15 Months of Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12–13

Well-Child Visits in the Third, Fourth, Fifth and Sixth Years of Life . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Adolescent Well-Child Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Prenatal Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Cesarean Section Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Asthma Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18–19

Hospitalizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Emergency Department Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20–21

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23–26

Changes in HEDIS Measures for Year 2011 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23




Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27–29

Summary of Medicaid/Commercial Comparison Studies Published Since 2000 . . . . . . . . . . . . . . . . 27–29

TABLE OF CONTENTS

2

EXECUTIVE SUMMARY

As part of the effort to assess and monitor Medicaid

performance under capitated managed care, the

Texas Health and Human Services Commission

(HHSC) contracts with the Institute for Child Health

Policy at the University of Florida as the External

Quality Review Organization (EQRO) to compare

access and quality measures of MCO performance to

state and national standards. To complement the

work of the EQRO, the University of Texas School of

Public Health (UTSPH) is using all payer administrative

claims and encounter data to compare access and

quality measures of Medicaid enrollees to populations

with commercial coverage.

FOR THIS STUDY, WE SELECTED FOUR MEASURES OF ACCESS TO PREVENTIVE CARE AND THREE MEASURES OF QUALITY OF CARE THAT ARE COMMONLY USED TO EVALUATE MCO PERFORMANCE FOR MEDICAID CHILDREN AND PREGNANT WOMEN:

• well child visits in the first 15 months of life,

• well child visits in the third, fourth, fifth and six years of life,

• adolescent well care visits,

• timeliness of prenatal care,

• Cesarean-section deliveries,

• childhood asthma hospitalizations, and

• childhood asthma emergency department visits.

We calculated Medicaid rates for these measures for

2011-2012 using Texas Medicaid enrollment, claims,

and encounter data and commercial rates for

2011-2013 using Truven’s Market Scan claims dataset.

We conducted a statistical test of the differences

between Medicaid MCO and commercial rates

and a logistic regression analysis controlling for

demographic, socioeconomic status (SES), and

geographic differences in the populations.

The results indicate that access and quality in the

Texas Medicaid program is comparable to that in the

commercial PPO insured population. Four out of

seven Medicaid MCO rates (well-child visits for

adolescents, timely prenatal care visits for pregnant

women, C-section deliveries, and inpatient stays for

asthmatic children) were better than PPO rates for

each comparison year, two were worse (well-child

visits for infants 0-15 months and ED visits for

asthmatic children), and one (well-child visits for

children aged 3-6 years) was better for Medicaid

MCO children in 2011 but worse in 2012 (see Table

below). Similarly, in the logistic regression analyses,

three out of seven odds ratios indicating the

likelihood of a well-child visit for adolescents, timely

prenatal care for pregnant women, and inpatient stay

for asthmatic children were significantly better for

Medicaid MCOs, three were better for commercial

PPO children (well-child visits for infants 0-15

months, well-child visits for children aged 3-6 years,

and ED visits for asthmatic children), and there

was no difference for one (the likelihood of a

C-section rate).

3

Medicaid MCO versus Commercial Performance

Rates Rates

Medicaid MCOEnrollees Truven PPO EnrolleesAdjusted Odds Ratio

(95% CI)2011 2012 2013 2011 2012 2013

Six or More Well-Child Visits Infants First 15 Months

61.1* 59.8* NA NA 82.8 93.8 .272 (.260 -. 284)

Well-Child Visits Age 3-6 77.8** 75.1* NA 77.5 81.0 NA .857 (.846 - .868)

Well-Child Visits Adolescents 60.8** 58.7** NA 40.8 43.0 NA 1.823 (1.805 – 1.842)

Timeliness of Prenatal Care 81.0** 75.6** NA 64.7 62.9 NA 2.225 (2.167 – 2.285)

Cesarean Delivery 34.3** 35.2** NA 39.7 39.4 NA .984 (.961 – 1.008)

Asthma Inpatient 122.8** 136.7** NA 383.4 387.8 NA .206 (.191 - .222)

Asthma ED Visits 1,201.0* 1,219.1* NA 771.2 775.2 NA 1.659 (1.550 – 1.775)

* Medicaid MCO rate significantly worse than Commercial PPO

** Medicaid MCO rate significantly better than Commercial PPO

NA – not available

EXECUTIVE SUMMARY

INTRODUCTION

Concerns about the rising cost of medical care and

the growing number of enrollees have led Texas

officials to turn to capitated managed care as the

predominant delivery model for Medicaid and

Children’s Health Insurance Program (CHIP) services.

Under this model, the state pays private managed

care organizations (MCOs) a fixed rate per member

per month, and the MCOs are responsible for the

cost of a defined set of services used by enrollees.

The MCOs may also offer value-added services such

as sports/gym memberships, respite care, pest

control, etc., to attract members. The evolution from

fee-for-service to the capitated MCO model for

Medicaid and CHIP enrollees began in the early

1990s and went through several periods of expansion

to different parts of the state and covered

populations. Currently, there are 19 MCOs serving 13

managed care service areas, with ten areas built

around major metropolitan population centers and

three covering large rural regions of the state.

In 2011, the percentage of the Medicaid population

enrolled in MCOs was 71%. As of November 2013,

81% of Medicaid enrollees were covered by MCOs

and the percentage was expected to exceed 90% by

the end of 2015.1 The two main Medicaid MCO

programs are State of Texas Access Reform (STAR),

which mainly covers acute care for pregnant women

and children, and STAR+PLUS, which combines

acute care with long-term services and supports for

adults and children who have a disability or who are

age 65 and older. STAR+PLUS also includes a large

population of beneficiaries who are dually enrolled in

both Medicaid and Medicare. The STAR Health

program which covers acute care for foster care

children and STAR Kids, which is being phased in this fall to

serve youth and children with disability-related Medicaid.

The federal government requires Texas to assess and

monitor Medicaid performance under capitated

managed care with respect to access and quality. As

part of the monitoring effort, the Health and Human

Services Commission (HHSC) contracts with the

Institute for Child Health Policy at the University of

Florida as the External Quality Review Organization

(EQRO) to compare utilization-based measures of

MCO performance to Medicaid state standards and

national benchmarks.2 To supplement the Medicaid

comparisons approach used by the EQRO for

assessing performance, the University of Texas

School of Public Health (UTSPH) faculty and students

are comparing utilization measures of access and

quality for Medicaid enrollees to that of similar

populations in Texas with commercial coverage using

all payer administrative claims and encounter data.

The specific objectives of the UTSPH study are to: 1)

compare recent trends in Health Effectiveness Data

and Information Set (HEDIS)-based healthcare

utilization rates among Medicaid and commercially

insured populations across the state, and 2)

determine at the individual level if access and quality

measures differ between Medicaid and commercial

enrollees after controlling for a robust set of

demographic, socio-economic, and geographic

covariates. This report provides background

information on access and quality activities of MCOs

and commercial plans, describes HEDIS measures

used in the study, and summarizes recent literature

on Medicaid/commercial performance comparisons

in other states. We then describe our data and

methods, and present results.

4

5

WELL-CHILD VISITS IN THE FIRST 15 MONTHS OF LIFE

The overall stated goal of Texas Medicaid MCOs is to

deliver quality, cost-effective care through the medical

home model. The MCO model is designed to ensure

that enrollees are connected to a primary care provider

and have access to a network of specialty providers.3

In addition, the MCOs participate in a number of

quality assessment and performance improvement

programs and reporting activities required under their

contracts with HHSC.3 MCOs participating in STAR,

STAR+Plus, CHIP, and STAR Health are required to

provide disease management programs covering

Asthma and Diabetes. STAR+Plus plans are also

required to provide these programs for chronic

obstructive pulmonary disease, congetstive heart

failure, and coronary artery disease. The MCOs are also

required to develop value-based purchasing initiatives

and must submit these plans to HHSC outlining the

proposed payment methods that encourage quality

outcomes and reduce inappropriate utilization of

services. In addition, the state’s contract requires each

STAR MCO to have a program for targeting,

outreaching, and educating members who have high

utilization patterns. HHSC has recently launched a

new initiative to conduct individual quarterly calls with

MCOs to maintain an interactive dialogue on priority

areas related to quality and efficiency.

Similar to Medicaid, commercial insurance plans are

experiencing increasing pressure to provide high

quality, lower cost care. The shift to “value-based” care

is a major objective of the Affordable Care Act of 2010.4

Some of the required changes for insurance carriers

under the ACA directly and indirectly impact access,

cost and quality. For example, beginning in 2010

consumers were able to go online to compare and

select amoung health insurance coverage options.

The lifetime cap on coverage benefits was eliminated

and many preventative services no longer require

copayments or cost sharing. Coverage has been

extended to a number of uninsured individuals with

pre-existing conditions as well as young adults who

became eligible under their parents’ insurance plan.

During 2011, large employer commercial plans were

required to spend 85 percent of their premium dollars

on healthcare service and/ or quality improvement.4

These commercial plans typically offer financial

incentives to providers who achieve specific

improvements in the health of members. The plans also

track health metrics in patients and providers may

receive financial rewards and recognition for meeting

the requirments of different quality improvement and

cost containment programs. Many commercial plans

are adopting quality improvement initiatives similar to

those in Medicare and Medicare Advantage plans. In

2012, the ACA established value-based payment

programs, which offer financial incentives to hospitals

to improve the quality of care and publicly reported

performance measures of network providers are being

put in place. In addition, incentives exist for physicians

to join and/or form Accountable Care Organizations

that incentivize doctors and hospitals to coordinate

patient care and improve quality. Implementing

alternative payment models such as bundled

payments, and other initiatives to build a more

competitive and transparent health insurance

marketplace are being piloted. Under the ACA, 2013

required essential health benefits for plans competing

in the small group and individual markets, but not the

large group market and in 2014 insurers were required

to provider coverage regardless of pre-existing

conditions. All told, because of the ACA and responses

to market pressures, access and quality have also been

a major emphasis of commercial plans in recent years.

BACKGROUND

6

BACKGROUND

HEDIS MEASURES OF ACCESS

AND QUALITY

The Healthcare Effectiveness Data and Information Set

(HEDIS) is a set of performance measures designed to

evaluate utilization and outcomes across manage care

plans. The reporting of HEDIS measures was originally

mandated by the Centers for Medicare and Medicaid

Services (CMS) for Medicare Advantage plans but has

since become an industry standard. Health plan

incentive payments are often tied to achieving HEDIS

benchmarks. Employers, consultants, and consumers

often use HEDIS data to help them select the best

health plan for their needs. Because HEDIS measures

are publicly available and are designed to influence

consumer behavior, managed care plans continue to

focus on achieving, improving and surpassing HEDIS

benchmark standards. Currently, HEDIS is used by

more than 90 percent of America’s health plans to

measure performance on important dimensions of care

and service. The HEDIS measures are updated annually

by NCQA’s Committee on Performance Measurement

to assure the measures continue to be relevant.6

Updates to HEDIS measures relevant for this study can

be found in the Appendix 1.

There are 83 HEDIS 2015 measures across five domains

of care. For this report, we selected four HEDIS

measures of access to preventive care that are relevent

to the Medicaid population:

• well child visits in the first 15 months of life,

• well child visits in the third, fourth, fifth and six years of life,

• adolescent well care visits, and

• timeliness of prenatal care (a component of the pre and post natal care measure).

In addition to the preventive care access measures, we

selected three population-relevent quality of care

indicators for the study:

• Cesarean-section delivery rates,

• childhood asthma hospitalization rates, and

• childhood asthma emergency department visit rates.

RELATED LITERATURE

We conducted a literature review to identify and

summarize studies published since 2000 that

compared similar access and quality indicators of

Medicaid and commerically insured populations in

other states. We searched the Pubmed database

using the following key words: Medicaid, commercial

or private insurance, claims, well child visits, prenatal

care, cesarean section, asthma, hospitalization, and

emergency care. After screening titles and abstracts,

we identified eight studies that reported and

compared utilization rates for well-child visits,

Cesarean section deliveries, and asthma care. The

table in Appendix 2 summarizes the key characteristics

of the studies and their major findings. Relevant

findings from the studies are summarized below.

WELL CHILD VISITS: Kenny and Coyer (2012), using

NHIS and MEPS survey data from 2008 and 2009,

reported comparable rates of primary and preventive

care among the two groups. Medicaid children

reported a slightly higher rate of well child visits while

having slightly lower rates of office visits compared to

those with commercial insurance. Ku (2009) also used

MEPS survey data from 2005 to show equivalency

between the average number of office visits and

outpatient hospital visits. A comparison of Medicaid

and commercially insured children conducted by the

7

U.S. Department of Health and Human Services

(Sebelius 2011) using claims data showed mixed

results for different age groups. Fifty-six percent of

Medicaid children 0-15 months of age had 6 or more

well child visits as compared to 76% of commercially

insured children. The gap between Medicaid and

commercially insured was smaller for older children,

and was reversed for adolescents. Sixty-four percent

of Medicaid children 3-6 years of age and 47% of

Medicaid children 12-21 years of age had at least one

well-child visit compared to 71% and 41% respectively

for the commercially insured population.

CESAREAN SECTION DELIVERIES: Only one study

was found comparing Cesarean section (C-section)

rates in Medicaid to commercially insured. It was

completed by linking Michigan medical record and

hospital discharge data by Movsas et al. (2012). The

study found that C-section rates were 29% of births for

Medicaid women and 33% for commercially insured

women during the timeframe of 2004 to 2008.

ASTHMA CARE: Merrick et al. (2001) found asthmatic

children on Medicaid had slightly longer lengths of

hospital stays compared to commercially insured

children. Medicaid children with asthma were more

likely to be re-admitted for hospitalization following a

hospital discharge for asthma. Peters et al (2008)

found that hospitalizations and ED visit rates of

Medicaid adults with severe or difficult-to-treat

asthma were significantly higher than commercially

insured (OR 3.08; 95% CI 2.11 - 4.50). Similarly,

Finkelstein et al (2000), in comparisons of asthma care

for Medicaid and commercially insured children within

the same HMO using 1991–1996 medical records and

claims data, found Medicaid children were more likely

to receive care in the ED (IRR 1.4; 95% CI 1.2 – 1.5) and

were more likely to be hospitalized for their asthma

(IRR1.3; 95% CI 1.1 – 1.5). Both groups had similar

rates of ambulatory visits for asthma.

Fingar and Washington (2015) reported a higher

percentage of potentially preventable pediatric

inpatient stays for asthma in Medicaid enrollees

compared to commercially insured populations using

the Nationwide Inpatient Sample (NIS) data during

2003-2012. In 2003, the two groups had almost

equivalent rates of 46.9% and 46.4%.. However, by

2012 the Medicaid rate had increased to 58.1% while

the rate for commercially insured children decreased

to 35.5%.

SUMMARY: The findings from the literature are

mixed. Three studies showed equivalency between

Medicaid and commercially insured based on rates for

well-child visits for children in different age groups.

One study suggested better performance for

Medicaid based on lower rates of C-section deliveries.

All four studies comparing hospitalization rates,

lengths of stays, and ED visits of Medicaid and

commercially insured children and adults with asthma

found higher rates for Medicaid.

BACKGROUND

8

METHODS

We selected a total of seven measures that summarize various aspects of access and quality among Medicaid and

commercially insured women and children. Three of these measures are well-child visit rates, two are pregnancy and

birth related rates, and two are asthma related rates. We calculated the rates for the entire Medicaid population and

for Medicaid MCO enrollees using the Texas Medicaid enrollment, claims, and encounter data obtained from the

Texas HHSC. We calculated commercial rates for PPO members using Truven’s Market Scan claims dataset. The

years of interest for both populations were 2011-2013 but due to administrative delays we have only been able to

include 2011 and 2012 data from Medicaid.

TRUVEN HEALTH ANALYTIC

MARKETSCAN COMMERCIAL CLAIMS

AND ENCOUNTERS

The Truven MarketScan database includes active

employees, early retirees, COBRA enrollees, and

dependents, all of whom are insured through

employer self-insured (ESI) plans. ESI plans are

regulated by the federal Employee Retirement

Income Security Act (ERISA). Employers who self-fund

their health plans pay the costs of their employee’s

health care themselves, rather than buying coverage

from an insurance company or HMO. Coverages may

vary by plan and employer. Employers who self-fund

their health plans may require employees to

contribute to the cost of the plan.

Federal law exempts large-group ESI plans from the

essential health benefits requirement of the ACA and

rating requirements that apply to individual and

small-group plans. Like other comprehensive health

plans, however, large-employer ESI plans must

provide preventive services with no copayments or

deductibles. Depending on age and gender, regular

check-ups, blood pressure and diabetes testing,

contraceptives, mammograms, cancer screenings,

and flu shots are covered free of copayments or

deductibles. The plans may not have lifetime or

annual dollar limits on coverage, and they can’t

deny coverage because of preexisting conditions

or health history.

Information on healthcare coverage eligibility and

individual level healthcare service use is included in

the Truven dataset. The data are divided into eight

tables which include: 1) inpatient admission records;

2) facility claims data; 3) individual facility and

professional encounters and services that the

inpatient admission record comprises; 4) outpatient

encounters and claims for services that were rendered

in a doctor’s office, hospital outpatient facility,

emergency room, or other outpatient facility; 5)

aggregated population tables which provide quarterly

counts of covered lives for medical/surgical and

outpatient pharmaceutical claims data for calculating

rate-based statistics; 6) outpatient pharmaceutical

claims; 7) annual enrollment summary tables;

and 8) person-level enrollment records with

demographic and plan information on users

and non-users of services.

The enrollment files contain information on age, birth

year, employee classification, employment status,

geographic location of employee, relation of enrollee

9

to employee, enrollment data, health plan type,

industry type, metropolitan statistical area (MSA),

gender, and three-digit zip code. The claims and

encounter files contain diagnosis codes (ICD 9 CM

diagnosis and DRG), procedure codes (ICD 9 CM

procedure, HCPCS, and CPT), revenue code, place of

service, provider type, along with information on type

of claim, date when claim was filed, dates of service

and payment information.

The commercial population selected for this study

were children 21 years of age and pregnant women

who were enrolled in PPO plans. These plans

represent the vast majority of individuals with

coverage in the Truven database. The study

population was linked to the medical claims and

encounter files and utilization rates were calculated

for each of the study’s outcome measures.

TEXAS MEDICAID CLAIMS

AND ENCOUNTERS

The Texas Medicaid data for 2011 and 2012 were

made available to UTSPH through an agreement with

HHSC. The Medicaid data consist of six files for each

year: 1) enrollment files with demographic, eligibility

and enrollment information in a per member per

month basis; 2) claim files containing the utilization

information for fee-for-service and primary care case

management Medicaid enrollees; 3) encounter files

containing utilization information for enrollees in

MCOs; 4) diagnosis files containing 25 additional

diagnosis fields apart from the primary diagnoses

contained in the claims files; 5) prescription drug files;

6) provider files with information about health care

providers who delivered the services. Texas Medicaid

rates were derived by linking the enrollment, claim

and encounter data for the study period. Medicaid

children < 21 years of age and pregnant women

whose enrollment was due to low-income status or

pregnancy were selected. Those enrolled due to

disability, foster care, social security income receipt,

legal or illegal immigrant status but requiring

emergency care, adoption subsidy receipt, or

dual-eligibility for Medicare and Medicaid

were excluded.

METHODS

10

METHODS

ACCESS AND QUALITY MEASURES

WELL-CHILD VISITS: As an indicator of access to

preventive services, the following well-child visit rates

were computed and compared for all continuously

enrolled Medicaid children, Medicaid children

enrolled in MCO plans, and children in Truven PPO

plans (Table 1):

• Children age 0 through 15 months with 6 or more well-child visits

• Children age 3 , 4, 5, or 6 with at least one well-child visit

• Adolescents age 12 – 21 years with at least one well-child visit

PRENATAL VISIT AND CESAREAN SECTION RATES: As indicators of acess and quality, the

following pregnancy related measures of utilizaiton

were computed for all continuously enrolled Medicaid

women, Medicaid women enrolled in MCO plans, and

women in Truven PPO plans (Table 1):

• Deliveries in which the mother received a prenatal care visit in the first trimester or within 42 days of enrollment

• Births by Cesarean section

ASTHMA CARE: The following asthma-related

utilization rates were computed for Medicaid children,

Medicaid children enrolled in MCO plans, and

children in Truven PPO plans (Table 1):

• Asthma related hospitalizations per 1,000 asthmatic children per year

• Asthma related emergency department visits per 1,000 asthmatic children per year

ANALYSIS

To assess performance based on rate comparisons,

we compared our seven measures of utilization for

Texas Medicaid enrollees with Texans who have

employer-based commercial PPO insurance using the

Truven database. We compared rates for each

measure for three covered populations for each year:

Medicaid, Medicaid MCO, and commercial PPO. To

assess whether the yearly rates between Medicaid and

the commercial PPO population were significantly

different, we completed a statistical test of the

differences in proportions between the Medicaid

MCO and commercial PPO population, given two

independent samples, using a z-score. Each outcome

measure is considered a proportion (e.g. the

Table 1. Measures Used to Assess Performance

Well-Child VisitsWell-Child Visits in the First 15 Months of Life

Well-Child Visits in the Third, Fourth, Fifth and Sixth Years of Life Adolescent Well-Care Visits

Pregnancy Related Outcomes Timeliness of Prenatal Care

Cesarean Section Deliveries

Asthma CareAsthma Inpatient Utilization

Asthma Emergency Department Utilization

11

proportion of children 0-15 months receiving well-

child visits), and the independent samples are the

Medicaid MCO and commercial PPO populations.

The null hypothesis is that a particular outcome

measure is the same for both populations, while the

alternative is that they are significantly different.

The test statistic used to test the hypothesis was :

where ^Pm is the estimated outcome measure in the

Medicaid MCO population, ^Pc is the estimated

outcome measure in the commercial PPO population, ^P is the estimated outcome measure in pooled

populations (Medicaid MCO and commercial PPO

together), Nm is the number of Medicaid MCO

enrollees, and Nc is the number of commercial PPO

enrollees. If the test statistic is greater than the upper

critical value (1.96), or lower than the lower critical

value (-1.96), then we rejected the null hypothesis at

the 5% significance level.

The second method used to test differences in each

measure across the two populations was a regression

analysis to determine the likelihood of a particular

outcome measure while controlling for underlying

demographic, socioeconomic status (SES), and

geographic differences in the populations that may

be correlated with the outcome measures. For this

method, we ran the following regression at the

individual enrollee level:

where Yi is the outcome measure of interest for

specific individuals in the Medicaid MCO and

commercial PPO population, Insi is a binary variable

that indicates whether the individual has Medicaid

MCO or commercial PPO insurance, Zi is a set of

control variables, including gender and age, SES

characteristics, and a geographic access indicator, and

Ei is uncontrolled error.

The main variable of interest is Insi, and the coefficient

on this variable tells us whether having commercial

insurance vis-à-vis Medicaid MCO coverage leads to

significantly better or worse outcomes, while

controlling for the remaining model variables. Gender

was coded as male or female and age a continuous

whole number variable for years old. Neither the

Medicaid nor Truven datasets contain individual-level

SES variables on employment, income, or education.

However, they contain zip code identifiers that were

used to obtain aggregate zip code-level SES variables

from the US Census data. Individuals within Medicaid

MCO and Truven PPO were given the SES and access

variables which corresponding to the zip code in

which they live. The SES zip-code level control

variables included: percent unemployed of the civilian

labor force age 16 years and over; percent of the

population with income in the past 12 months below

the federal poverty level; median rent in dollars for

occupied units paying rent; and percent of the

population with a bachelor’s degree or higher. The

geographic access control variable was the number of

primary care physicians per 100,000 population within

the member’s zip code area.

10

populations. The null hypothesis is that a particular outcome measure is the same for both populations, while the alternative is that they are significantly different. The test statistic used to test the hypothesis was :

𝑍𝑍 = �̂�𝑝𝑚𝑚 − �̂�𝑝𝑐𝑐

√�̂�𝑝(1 − �̂�𝑝) ( 1𝑛𝑛𝑚𝑚

+ 1𝑛𝑛𝑐𝑐

)

where �̂�𝑝𝑚𝑚 is the estimated outcome measure in the Medicaid MCO population, �̂�𝑝𝑐𝑐 is the estimated outcome measure in the commercial PPO population, 𝑝𝑝 ̂is the estimated outcome measure in pooled populations (Medicaid MCO and commercial PPO together), 𝑛𝑛𝑚𝑚 is the number of Medicaid MCO enrollees, and 𝑛𝑛𝑐𝑐 is the number of commercial PPO enrollees. If the test statistic is greater than the upper critical value (1.96), or lower than the lower critical value (-1.96), then we rejected the null hypothesis at the 5% significance level.

The second method used to test differences in each measure across the two populations was a regression analysis to determine the likelihood of a particular outcome measure while controlling for underlying demographic, socioeconomic status (SES), and geographic differences in the populations that may be correlated with the outcome measures. For this method, we ran the following regression at the individual enrollee level:

𝑦𝑦𝑖𝑖 = 𝛼𝛼 + 𝛽𝛽1𝐼𝐼𝑛𝑛𝐼𝐼𝑖𝑖 + 𝛾𝛾𝑍𝑍𝑖𝑖 + 𝜀𝜀𝑖𝑖 where 𝑦𝑦𝑖𝑖 is the outcome measure of interest for specific individuals in the Medicaid MCO and commercial PPO population, 𝐼𝐼𝑛𝑛𝐼𝐼𝑖𝑖 is a binary variable that indicates whether the individual has Medicaid MCO or commercial PPO insurance, 𝑍𝑍𝑖𝑖 is a set of control variables, including gender and age, SES characteristics, and a geographic access indicator, and 𝜀𝜀𝑖𝑖 is uncontrolled error.

The main variable of interest is 𝐼𝐼𝑛𝑛𝐼𝐼𝑖𝑖, and the coefficient on this variable tells us whether having commercial insurance vis-à-vis Medicaid MCO coverage leads to significantly better or worse outcomes, while controlling for the remaining model variables. Gender was coded as male or female and age a continuous whole number variable for years old. Neither the Medicaid nor Truven datasets contain individual-level SES variables on employment, income, or education. However, they contain zip code identifiers that were used to obtain aggregate zip code-level SES variables from the US Census data. Individuals within Medicaid MCO and Truven PPO were given the SES and access variables which corresponding to the zip code in which they live. The SES zip-code level control variables included: percent unemployed of the civilian labor force age 16 years and over; percent of the population with income in the past 12 months below the federal poverty level; median rent in dollars for occupied units paying rent; and percent of the population with a bachelor’s degree or higher. The geographic access control variable was the number of primary care physicians per 100,000 population within the member’s zip code area.

10

populations. The null hypothesis is that a particular outcome measure is the same for both populations, while the alternative is that they are significantly different. The test statistic used to test the hypothesis was :

𝑍𝑍 = �̂�𝑝𝑚𝑚 − �̂�𝑝𝑐𝑐

√�̂�𝑝(1 − �̂�𝑝) ( 1𝑛𝑛𝑚𝑚

+ 1𝑛𝑛𝑐𝑐

)

where �̂�𝑝𝑚𝑚 is the estimated outcome measure in the Medicaid MCO population, �̂�𝑝𝑐𝑐 is the estimated outcome measure in the commercial PPO population, 𝑝𝑝 ̂is the estimated outcome measure in pooled populations (Medicaid MCO and commercial PPO together), 𝑛𝑛𝑚𝑚 is the number of Medicaid MCO enrollees, and 𝑛𝑛𝑐𝑐 is the number of commercial PPO enrollees. If the test statistic is greater than the upper critical value (1.96), or lower than the lower critical value (-1.96), then we rejected the null hypothesis at the 5% significance level.

The second method used to test differences in each measure across the two populations was a regression analysis to determine the likelihood of a particular outcome measure while controlling for underlying demographic, socioeconomic status (SES), and geographic differences in the populations that may be correlated with the outcome measures. For this method, we ran the following regression at the individual enrollee level:

𝑦𝑦𝑖𝑖 = 𝛼𝛼 + 𝛽𝛽1𝐼𝐼𝑛𝑛𝐼𝐼𝑖𝑖 + 𝛾𝛾𝑍𝑍𝑖𝑖 + 𝜀𝜀𝑖𝑖 where 𝑦𝑦𝑖𝑖 is the outcome measure of interest for specific individuals in the Medicaid MCO and commercial PPO population, 𝐼𝐼𝑛𝑛𝐼𝐼𝑖𝑖 is a binary variable that indicates whether the individual has Medicaid MCO or commercial PPO insurance, 𝑍𝑍𝑖𝑖 is a set of control variables, including gender and age, SES characteristics, and a geographic access indicator, and 𝜀𝜀𝑖𝑖 is uncontrolled error.

The main variable of interest is 𝐼𝐼𝑛𝑛𝐼𝐼𝑖𝑖, and the coefficient on this variable tells us whether having commercial insurance vis-à-vis Medicaid MCO coverage leads to significantly better or worse outcomes, while controlling for the remaining model variables. Gender was coded as male or female and age a continuous whole number variable for years old. Neither the Medicaid nor Truven datasets contain individual-level SES variables on employment, income, or education. However, they contain zip code identifiers that were used to obtain aggregate zip code-level SES variables from the US Census data. Individuals within Medicaid MCO and Truven PPO were given the SES and access variables which corresponding to the zip code in which they live. The SES zip-code level control variables included: percent unemployed of the civilian labor force age 16 years and over; percent of the population with income in the past 12 months below the federal poverty level; median rent in dollars for occupied units paying rent; and percent of the population with a bachelor’s degree or higher. The geographic access control variable was the number of primary care physicians per 100,000 population within the member’s zip code area.

METHODS

12

RESULTS

WELL-CHILD VISITS IN THE FIRST 15 MONTHS OF LIFE

We calculated the rates of well-child visits in the first 15 months of life for continuously enrolled Medicaid children

and Medicaid MCO children in 2011 and 2012. We calculated comparative rates for commercial PPO children in

2012 and 2013.

Medicaid children had lower rates compared to commercial PPO insured children and all Medicaid children had

lower rates compared to Medicaid MCO children (Table 2). The 2012 Medicaid MCO rate was statistically

significant and lower than the commercial PPO rate. It is noteworthy that the rate for commercial PPO insured

children in Texas is also higher than the HEDIS national PPO rate, which was 76.1% in 2011 and 76.4% in 2012.

Further evaluation shows nearly all (99%) of Medicaid MCO children receive at least one visit and the majority

(over 88%) receive four or more visits (Table 3).

Table 2: Well-Child Visits in the First 15 Months of Life (6 or More Visits)

Year Medicaid Medicaid MCO Truven PPO

Eligible Population Percent Eligible

Population Percent Eligible Population Percent

2011 137,936 28.28% 49,989 61.08% - -

2012 119,985* 26.43% 41,379 59.77% 14,902 82.75%

2013 - - - - 6,185 93.79%

*The 2012 population for Texas Medicaid contains data only for the first 8 months of the calendar year.

Table 3: Well-Child Visits in the First 15 Months of Life (Number of Visits)

Year Insurance Eligible Population 0 Visits 1 Visits 2 Visits 3 Visits 4 Visits 5 Visits 6 Visits

2011Medicaid 137,936 35.47% 2.96% 4.39% 7.26% 10.47% 11.16% 28.28%Medicaid 49,989 1.00% 1.71% 2.97% 5.61% 10.25% 17.38% 61.08%MCO

2012

Medicaid 119,985 20.50% 9.96% 8.42% 9.53% 12.99% 12.16% 26.43%Medicaid 41,379 1.18% 1.74% 3.02% 5.45% 10.47% 18.37% 59.77%MCOTruven 14,902 8.48% 0.40% 0.74% 1.091% 1.81% 4.723% 82.75%PPO

2013Truven 6,185 0.37% 0.29% 0.50% 0.78% 1.39% 2.88% 93.79%PPO


13

After controlling for demographic, SES, and gographic access differences, the likelihood of a Medicaid MCO

infant having six visits from age 0-15 months was only 27% of a commercial PPO insured infant (Table 4).

Table 4: Well-Child Visits in the First 15 Months of Life Regression Results

Six or More Visits Odds Ratios [95% CI]MedicaidMCO* 0.272 0.260 0.284

Percent Female 1.008 0.982 1.035

Percent Unemployed 0.997 0.988 1.006

Percent Below Poverty* 0.993 0.992 0.995

Median Rent 1.000 1.000 1.000

Percent with Bachelor's Degree* 1.004 1.002 1.005

Primary Care Providers per 100,000 1.000 1.000 1.000

Constant 6.155 5.650 6.707

Number of obs = 105284

Prob > chi2 = 0.0000

Pseudo R2 = 0.0418

RESULTS

14

RESULTS

WELL-CHILD VISITS IN THE THIRD, FOURTH, FIFTH AND SIXTH YEARS OF LIFE:

We calculated the rates of well-child visits in years three through six for Medicaid children and Medicaid MCO

children in 2011 and 2012. We calculated comparative rates for commercial PPO children in 2012 and 2013.

Medicaid MCO rates for well-child visits in years three through six are slightly lower in 2012 than for commericial

PPO insured children, but substantially higher than the Medicaid population (Table 5). Medicaid rates showed

significant improvement between 2011 and 2012, but remained significantly lower than the rates of Medicaid

MCO and commercial PPO insured children. The Medicaid MCO rate was statistically significant, higher than the

commercial PPO rate in 2011, and lower than the commercial PPO rate in 2012.

Table 5: Well-Child Visits in the Third, Fourth, Fifth and Sixth Years of Life (Percent Children with at Least 1 Visit)

Year Medicaid Medicaid MCO Truven PPOEligible

Population Percent Eligible Population Percent Eligible

Population Percent

2011 554,692 48.85% 255,651 77.82% 157,771 77.53%

2012 585,827 63.51% 309,492 75.06% 162,229 81.04%

* Texas Medicaid data is based on fiscal year and Truven data is based on calendar year.

After controlling for demographic, SES, and geographic access differences, the likelihood of a Medicaid MCO

child age three, four, five, or six having a well-child visit was 86% of a commercial PPO insured child (Table 6).

Table 6: Well-Child Visits in the Third, Fourth, Fifth and Sixth Years of Life Regression Results

Well-Child Visits Odds Ratios [95% CI]MedicaidMCO* 0.857 0.846 0.868

Percent Female* 0.976 0.966 0.986

Age* 0.737 0.734 0.741

Percent Unemployed* 1.041 1.038 1.045


Median Rent 1.000 1.000 1.000



Constant 8.084 7.775 8.405


Prob > chi2 = 0.0000

Pseudo R2 = 0.0216

15

ADOLESCENT WELL-CHILD VISITS

The percentage of Medicaid MCO adolescents between the ages of 12 and 21 who had at least one well-child

visit is higher than the percentage for commercial PPO insured and for all Medicaid adolescents (Table 7).

Medicaid rates increased between 2011 and 2012, and in 2012 were higher than commercial PPO insured. The

Medicaid MCO rate was statistically significant and higher than the commercial PPO rate in both years.

Table 7: Adolescent Well-Child Visits (% of Children with at Least 1 Visit)



Population Percent

2011 418,938 35.71% 180,636 60.79% 446,449 40.79%

2012 459,616 49.54% 235,286 58.67% 501,613 42.96%



adolescent having a well-child visit was 1.8 times greater than a commercial PPO insured adolescent (Table 8).

Table 8: Adolescent Well-Child Visits Regression Result

Well-Child Visits Odds Ratios [95% CI]MedicaidMCO* 1.823 1.805 1.842

Percent Female* 1.392 1.382 1.402

Age (years)* 0.848 0.847 0.849



Median Rent 1.000 1.000 1.000



Constant 3.047 2.947 3.150


Prob > chi2 = 0.0000

Pseudo R2 = 0.0618

RESULTS

16

RESULTS

PRENATAL VISITS

Texas Medicaid MCO rates for the the percent of deliveries with a timely prenatal care visit exceeded those of

both commercial PPO insured and all Medicaid (Table 9). The Medicaid MCO rate was statistically significant and

higher than the commercial PPO rate in both years.

Table 9: Percent of Deliveries with Timely Prenatal Care



Population Percent

2011 144,338 57.83% 85,479 81.04% 9,302 64.72%

2012 96,039 54.52% 76,506 75.62% 43,530 62.91%


pregnant woman having a timley prenatal care visit was 2.2 times greater than for a commercial PPO insured

pregnant woman (Table 10).

Table 10: Percent of Deliveries with Timely Prenatal Care Regression Results

Timely Prenatal Care Odds Ratios [95% CI]MedicaidMCO* 2.225 2.167 2.285

Age 20-24* 1.293 1.247 1.340

Age 25-29* 1.302 1.255 1.352

Age 30-34* 1.394 1.340 1.452

Age 35+* 1.610 1.537 1.686



Median Rent 1.000 1.000 1.000



Constant 1.066 0.993 1.144


Prob > chi2 = 0.0000

Pseudo R2 = 0.0226

17

CESAREAN SECTION RATES

The percentage of Cesarean section deliveries was higher for commercial PPO insured women than Medicaid

MCO and all Medicaid women (Table 11). The Medicaid MCO rate was statistically significant and lower than the

commercial PPO rate in both years.

Table 11: Cesarean Section Delivery


DeliveriesPercent of C-section

Eligible Deliveries

Percent of C-section

Eligible Deliveries

Percent of C-section

2011 94,584 35.49% 52,391 34.32% 41,274 39.72%

2012 59,065 35.20% 57,718 35.18% 45,342 39.39%



woman having a C-section delivery was not significantly different from a commercial PPO insured woman (Table 12).

Table 12: Cesarean Section Delivery Regression Result

Cesarean Section Delivery Odds Ratios [95% CI]MedicaidMCO* 0.984 0.961 1.008

Age 20-24* 1.425 1.370 1.482

Age 25-29* 1.798 1.728 1.871

Age 30-34* 2.195 2.106 2.287

Age 35+* 3.029 2.895 3.168



Median Rent 1.000 1.000 1.000



Constant 0.326 0.304 0.350


Prob > chi2 = 0.0000

Pseudo R2 = 0.0153

RESULTS

18

RESULTS

ASTHMA CARE

HOSPITALIZATIONS: The Texas Medicaid MCO asthmatic population has lower rates of inpatient stays per 1,000

asthmatic children per year compared to the Texas commercial PPO insured population for 2011 and 2012 (Table

13). This is also true comparing the Medicaid MCO to the entire Medicaid population. The Medicaid MCO rate

was statistically significant and lower than the commercial PPO rate in both years.

Table 13: Asthma Inpatient Stays Among Asthmatic Children(Inpatient Stays/1,000 Member Years)


PopulationHospitalization

RateEligible


RateEligible


Rate

2011 39,397 104.47 25,242 122.82 1,985 383.38

2012 41,613 107.73 40,622 136.65 2,484 387.84



asthmatic child having a hospitalization was only 21% as likely as a commercial PPO insured child (Table 14).

Table 14: Asthma Inpatient Stays Among Asthmatic Children Regression Results

Asthma Inpatient Odds Ratios [95% CI]MedicaidMCO* 0.206 0.191 0.222

Percent Female 0.954 0.905 1.005

Age (years)* 0.941 0.935 0.947



Median Rent ($) 1.000 1.000 1.000

Percent with Bachelor's Degree 0.999 0.996 1.002


Constant 1.082 0.920 1.273


Prob > chi2 = 0.0000

Pseudo R2 = 0.0448

19

EMERGENCY DEPARTMENT VISITS: ED visits for enrollees identified as asthmatics were lower for commercial

PPO insured children compared to both Texas Medicaid and Medicaid MCO enrollees (Table 15). The rate for all

Medicaid was lower than Medicaid MCO. The rates for commercial PPO insured and Medicaid MCO children were

stable over the two years, while the Medicaid rate increased. The Medicaid MCO rate was statistically significant

and higher than the commercial PPO rate in both years.

Table 15: Asthma ED Visits Among Asthmatic Children(ED Visits/1,000 Member Years)


Population ED Visit Rate Eligible Population ED Visit Rate Eligible

Population ED Visit Rate

2011 39,397 873.81 25,242 1,201.02 1,985 771.21

2012 41,613 958.53 40,622 1,219.11 2,484 775.24



child having an ED visit was 1.7 times greater than a commercial PPO insured child (Table 16).

Table 16: Asthma ED Visits Among Asthmatic Children Regression Results

Asthma ED Odds Ratios [95% CI]MedicaidMCO* 1.659 1.550 1.775

Percent Female* 0.916 0.884 0.949

Age (years)* 1.016 1.011 1.020



Median Rent ($)* 1.000 1.000 1.000

Percent with Bachelor's Degree 0.999 0.997 1.001

Primary Care Providers per 100,000* 1.001 1.001 1.001

Constant 1.440 1.278 1.621


Prob > chi2 = 0.0000

Pseudo R2 = 0.016

RESULTS

20

CONCLUSION

We compared seven measures of access and quality

for children and pregnant women living in Texas with

Medicaid MCO coverage versus commercial PPO

insurance on both aggregate rates and at the

individual level controlling for demographic, SES, and

geographic access. The results provide mixed

evidence on how well the Texas Medicaid MCO

program is doing in closing the gap in access and

quality of care for this population compared to the

commercially covered PPO populations.

Four out of seven aggregate rates (well-child visits for

adolescents, timely prenatal care visits for pregnant

women, C-section rates of deliveries, and inpatient stays

for asthmatic children) are statistically significant and

better for Medicaid MCOs for each comparison year

while three of the outcome measures are better for

commercial PPO insured children (well-child visits for

infants 0-15 months, well-child visits for children aged 3-6

years, and ED visits for asthmatic children) (Table 17).

Similarly, in the individual level analyses controlling for

population differences in demographic, SES, and

geographic access, three out of seven odds ratios

(well-child visits for adolescents, timely prenatal care

visits for pregnant women, and inpatient stays for

asthmatic children) were statistically significant and

better for Medicaid MCOs, while three were better for

commercial PPO insured children (well-child visits for

infants 0-15 months, well-child visits for children aged

3-6 years, and ED visits for asthmatic children), and

there was no difference for one indicator (the

likelihood of a C-section rate) (Table 17).

Table 17: Medicaid versus Commercial Performance Scorecard

Rates Rates

Medicaid MCOEnrollees Truven PPO EnrolleesAdjusted Odds Ratio

(95% CI)2011 2012 2013 2011 2012 2013

Well-Child Visits Infants 61.1* 59.8* NA NA 82.8 93.8 .272 (.260 -. 284)

Well-Child Visits Age 3-6 77.8** 75.1* NA 77.5 81.0 NA .857 (.846 - .868)

Well-Child Visits Adolescents 60.8** 58.7** NA 40.8 43.0 NA 1.823 (1.805 – 1.842)

Timeliness of Prenatal Care 81.0** 75.6** NA 64.7 62.9 NA 2.225 (2.167 – 2.285)

Cesarean Delivery 34.3** 35.2** NA 39.7 39.4 NA .984 (.961 – 1.008)

Asthma Inpatient 122.8** 136.7** NA 383.4 387.8 NA .206 (.191 - .222)

Asthma ED Visits 1,201.0* 1,219.1* NA 771.2 775.2 NA 1.659 (1.550 – 1.775)

* Medicaid MCO rate significantly worse than Commercial PPO

** Medicaid MCO rate significantly better than Commercial PPO

NA – not available

21

The individual level analyses also indicate particularly

large differences in the likelihood ratios for the MCO

population compared to commercial PPO insured on

certain measures. Infants were only .27 as likely to

have six visits over first 15 months. Pregnant women 2.2

times more likely to have timely prenatal care. Asthmatic

children were only .21 as likely to have a hospitalization

but 1.7 times more likely to have an ED visit.

Compared to the literature, our findings suggest

Texas Medicaid MCOs may be doing better in asthma

care for children as indicated by lower hospitalization

rates compared to commercial PPO children.

However, like the national comparisons, ED visit rates

for Medicaid MCO children with asthma were

substantially higher. Texas Medicaid MCOs are not

doing as well on well-child visits for infants and young

children. Yet, the literature indicates roughly

equivalent rates of access to well-child visits between

Medicaid and commercially insured children in all age

groups in other States. In Texas, infants and young

children in Medicaid MCOs had significantly lower

rates of well-child visits than commercial PPO insured

children. However, it is noteworthy that the majority

of Texas Medicaid MCO infants had more than four

visits, indicating the gap primarily lies in achieving

more than four visits. Medicaid adolescents enrolled

in MCOs had higher rates of receiving a well-child visit

than their commercial PPO counterparts. Texas

Medicaid MCO outcomes for births, measured both

by timeliness of prenatal care and C-section rates,

were also better than commercial PPO rates globally;

however, the individual level analyses indicated no

difference. This finding is consistent with one other

study found in the literature.

In conclusion, the aggregate and individual level

comparisons suggest that access and quality in the

Texas Medicaid program is comparable to that in the

commercial PPO insured population, e.g, it is doing

better on some measures and worse on others. This

conclusion is based on both aggregate rate

comparisons and on individual likelihood analyses

controlling for demographic, SES, and geographic

access differences between the populations.

CONCLUSION

22

REFERENCES

1. Texas Health and Human Services Commission, Texas Medicaid and CHIP in Perspective, 10th Edition, February 2015.

2. The Institute for Child Health Policy, University of Florida, External Quality Review Organization, Summary of Activities and Trends in Healthcare Quality, Contract Year 2014.

3. Texas Health and Human Services Commission. Quality Based Payment and Delivery Reforms in Medicaid and the Children’s Health Insurance Program. February 2016.

4. U.S. Department of Health and Human Services. Key Features of the Affordable Care Act By Year. Retrieved August 11, 2016, from http://www.hhs.gov/healthcare/facts-and-features/key-features-of-aca-by-year/index.html#2011

5. Koonsman, G. Changes in Commercial Insurance Coverage & The Impact on Health System Value. 2016 March 31. Retrieved August 11, 2016, from http://www.hfma.org/uploadedFiles/Education/Capital_Conference/VMG%20Health%20Presentation_HFMA%20Capital.pdf

6. HEDIS® and Quality Compass®. (n.d.). Retrieved August 11, 2016, from http://www.ncqa.org/hedis-quality-measurement/what-is-hedis

7. Kenney GM and Coyer C. National Findings on Access to Health Care and Service Use for Children Enrolled in Medicaid or CHIP. MACPAC Contractor Report No. 1. Urban Institute. March 2012.

8. Ku L. Medical and Dental Care Utilization and Expenditures under Medicaid and Private Health Insurance. Med Care Res Rev. 2009 Aug;66(4):456-71. doi: 10.1177/1077558709334896. Epub 2009 Apr 23.

9. Sebelius K. Annual Report on the Quality of Care for Children in Medicaid and CHIP, U.S. Department of Health and Human Services, December 2011.

10. Movsas TZ, Wells E, Mongoven A, Grigorescu V. Does Medical Insurance Type (Private vs Public) Influence the Physician's Decision to Perform Caesarean Delivery? J Med Ethics. 2012 Aug;38(8):470-3. doi: 10.1136/medethics-2011-100209. Epub 2012 May 5.

11. Merrick N, Houchens R, Tillisch S, and Berlow B. Quality of Hospital Care of Children with Asthma: Medicaid Versus Privately Insured Patients. Journal of Health Care for the Poor and Underserved, Vol. 12, No. 2 (2001), pp. 192–207.

12. Peters AT, Klemens JC, Haselkorn T, Weiss ST, Grammer LC, Lee JH, Chen H. TENOR Study Group. Insurance Status and Asthma-related Health Care Utilization in Patients with Severe Asthma. Ann Allergy Asthma Immunol. 2008 Apr;100(4):301-7. doi: 10.1016/S1081-1206(10)60590-X.

13. Finkelstein JA, Barton MB, Donahue JG, Algatt-Bergstrom P, Markson LE, Platt R. Comparing Asthma Care for Medicaid and Non-Medicaid Children in a Health Maintenance Organization. Arch Pediatr Adolesc Med. 2000 Jun;154(6):563-8.

14. Fingar K, Washington R. Potentially Preventable Pediatric Hospital Inpatient Stays for Asthma and Diabetes, 2003–2012: Statistical Brief #192. Healthcare Cost and Utilization Project.

23

APPENDIX 1

CHANGES IN RELEVANT HEDIS MEASURES 2011-2014

Because the HEDIS measures are a focus for managed care plans, it is plausible that managed care organizations

adjust resources to assure their achievement. Each year managed care organizations assess their HEDIS

performance and may adjust benefit designs and/ or resources, as needed. Table 1 – 4 summarizes changes in the

relevant HEDIS measures between 2011 and 2014.

Changes in HEDIS Measures for Year 2011

HEDIS Measure Commercial Medicaid Modifications

Well-Child Visits in the First 15

Months of Life Yes Yes

Added ICD-9-CM Diagnosis code

V20.3 to Table W15-A.

Well-Child Visits in the Third,

Fourth, Fifth and Sixth Years of LifeYes Yes No changes to this measure.

Adolescent Well-Care Visits Yes Yes No changes to this measure.

Prenatal and Postpartum Care Yes Yes

• Clarified step 1 in the Administrative Specification.

• Added CPT code 99500 to Table PPC-C (Decision Rules 2, 3 and 4).

• Added LOINC codes 56990-5, 56991-3, 57321-2, 57743-7 to Table PPC-C (Decision Rules 2 and 3).

• Added CPT code 99500 to Table PPC-D.

• Added CPT code 99501 to Table PPC-E.

• Clarified that ultrasounds and lab results alone should not be considered a visit in the Note section.

• Added a practitioner type requirement to the Postpartum Care numerator.

24

APPENDIX 1



Well-Child Visits in the First

15 Months of Life Yes Yes HCPCS codes G0438, G0439 t.

Well-Child Visits in the

Third, Fourth, Fifth and Sixth

Years of Life

Yes Yes HCPCS codes G0438, G0439 t

Adolescent Well-Care Visits Yes YesAdded HCPCS codes G0438, G0439 to Table

AWC-A

Prenatal and Postpartum

CareYes Yes

• Moved CPT code 99500 from Part A to Part D in Table PPC-C Decision Rule 2.

• Added LOINC codes 972-0, 978-7, 1305-2 to Table PPC-C (Decision Rules 2 and 3).

• Deleted CPT code 99500 and HCPCS codes H1000-H1005 from Table PPC-C Decision Rule 3; this eliminates redundancy because members with these codes are identified in Decision Rule 4.

• Moved CPT code 99500 from Part A to Part C in Table PPC-C Decision Rule 4.

• Moved CPT code 99500 from Part C to Part A in Table PPC-D.

• Clarified in the Note section that the most recent estimated date of delivery (EDD) should be used if multiple dates are documented and that a single date (date of delivery or EDD) must be used to define the start and end of the first trimester.

• Clarified in the Note section that postpartum visits with physician assistants, nurse practitioners, midwives and registered nurses are acceptable.

• Clarified in the Note section that the intent of the measure is to assess whether prenatal and preventive care was rendered on a routine, outpatient basis.

25




15 Months of Life Yes Yes

• Revised example in continuous enrollment to account for leap year.

• Deleted obsolete CPT code 99432 from Table W15-A.



Years of Life

Yes Yes No changes to this measure

Adolescent Well-Care Visits Yes Yes No changes to this measure


CareYes Yes

Clarified in the Note section that the organization must define a method to determine which EDD to use and use one date consistently if multiple dates are documented.

APPENDIX 1

26




15 Months of Life Yes Yes

• Removed coding tables and replaced all coding table references with value set references.

• Revised example in continuous enrollment to account for a year that is not a leap year.

• Clarified that visits must be on different dates of services for the numerators in the Administrative Specification



Years of Life

Yes YesRemoved coding tables and replaced all coding

table references with value set references.

Adolescent Well-Care Visits Yes YesRemoved coding tables and replaced all coding

table references with value set references.


CareYes Yes

• Removed coding tables and replaced all coding table references with value set references.

• Removed Definition section.

• Moved steps to identify the eligible population (previously steps 1 and 2 under the Denominator section in the Administrative specification) to the Eligible Population section.

• Removed references to —family practitioner and —midwife because these practitioners are included in the definitions of PCP and OB/GYN and other prenatal care practitioners, respectively.

• Consolidated the steps for identifying numerator events.

• Consolidated four decision rules (formerly in Table

PPC-C) into three decision rules.

APPENDIX 1

27

Summary of Medicaid/Commercial Comparison Studies Published Since 2000

Indicator Publication Data Source and Year

Comparison Groups Measure Results

Well Child Visits

Kenney and Coyer (2012)

- National Health Interview Survey (NHIS)

- Medical Expenditure Panel Survey (MEPS) - 2008-09

- Medicaid children

- Children with employer sponsored commercial insurance

Well-child visit in previous 12 months

Medicaid 81.7%; ESI 81.6% *

Office visit in previous 12 months

Medicaid 93.9%; ESI 94.5% *

Usual source of care

Medicaid 95.5%; ESI 97.3%

Ku (2009) MEPS – 2005 Low-income children 0-18

- Medicaid children

- Children with commercial insurance

Annual number of office visits

Medicaid 2.49; Commercial 2.87 (p > 0.05)

Annual number of hospital outpatient visits

Medicaid .10; Commercial .18 (p > 0.05)

Sebelius (2011)

- CMS Medicaid claims data

- National Committee for Quality Assurance (NCQA) report - 2010

- Medicaid/ CHIP children


6+ well child visits, 0-15 months

Medicaid/CHIP 56%; Commercial 76% *

1+ well child visits, 3 – 6 years


1+ well child visits, 12 – 21 years


Cesarean section deliveries

Movsas et al (2012)

- Michigan Vital Records

- Michigan Inpatient Hospital Database - 2004-08

- Medicaid births

- Births of women with commercial insurance

C-section rate Medicaid 29%; Private 33% *

C-section odds ratio

Commercial vs. Medicaid 1.20, 95% CI (1.19 - 1.22)

*Results of statistical significance not provided

APPENDIX 2

28

APPENDIX 2




Asthma

Peters et al (2008)

Multicenter study interview questionnaire -2001-04

-Medicaid adults wih severe or difficult-to-treat asthma

- Adults with severe or difficult-to-treat asthma with commercial insurance

Hospitalization for asthma

Medicaid 18%; Commercial 7% (p < 0.001)

ED visits for asthma


Hospitalization or ED visits for asthma


Hospitalization or ED visits for asthmatics, odds ratio

Medicaid vs Commercial 3.08, 95% CI (2.11- 4.50)

Merrick et al

(2001)

Medical records for California, Georgia and Michigan children -1991

Children - Medicaid - All Commercial

Average LOS for asthma (days)

California: Medicaid 2.87; Commercial 2.27 (p 0.001)

Georgia: Medicaid 3.14; Commercial 2.80 (p 0.007)

Michigan: Medicaid 2.81; Commercial 2.63 (p 0.101)

Prior admission within 2 weeks for asthma

California: Medicaid 2.3%; Commercial 0.8% (p 0.083)

Georgia:

Medicaid 1.6%; Commercial 1.4% (p 0.803)

Michigan: Medicaid 3.8%; Commercial 2.0 (p 0.113)


29




Asthma

Rate of ED visits Medicaid vs Commercial IRR 1.4; 95% CI

(1.2 – 1.5)

Rates of Medicaid vs.

Fingar and Washington (2015)

Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS) -2003-11

- Medicaid children


% potentially preventable pediatric inpatient stays for asthma

2003: Medicaid 46.9%; Commercial 46.4%

2012: Medicaid 58.1%; Commercial 35.5% *


APPENDIX 2

Corporate Headquarters4507 Medical Dr., San Antonio, TX 78229

MHM.org

Wesley Health & Wellness Center

1406 Fitch Street, San Antonio, TX 78211

Bishop Ernest T. Dixon, Jr. Clinic

1954 E. Houston St., Ste. 201, San Antonio, TX 78202

School Based Health Center at Krueger Elementary

217 West Otto Street, Marion, TX 78124

School Based Health Center at Schertz Elementary

757 Curtiss Ave., Schertz, TX 78154

Methodist Healthcare Ministries of South Texas, Inc. is a private, faith-based,

not-for-profit organization dedicated to creating access to health care for

uninsured and low-income families through programs and services, strategic

grant-making and advocacy in 74 counties across South Texas.

The mission of Methodist Healthcare Ministries is “Serving Humanity to Honor

God” by improving the physical, mental and spiritual health of those least served

in the Rio Texas Conference area of The United Methodist Church.

The mission also includes Methodist Healthcare Ministries’ one-half ownership of

the Methodist Healthcare System – the largest healthcare system in South Texas.

This creates a unique avenue to ensure that Methodist Healthcare System

continues to be a benefit to the community by providing quality care to all and

charitable care when needed.

TEXAS - cdn.ymaws.com

Documents