Clinical Metric and Medication Persistency Effects: Evidence from a Medicaid Care Management Program

Abstract

The objective was to estimate clinical metric and medication persistency impacts of a care management program. The data sources were Medicaid administrative claims for a sample population of 32,334 noninstitutionalized Medicaid-only aged, blind, or disabled patients with diagnosed conditions of asthma, coronary artery disease, chronic obstructive pulmonary disease, diabetes, or heart failure between 2005 and 2009. Multivariate regression analysis was used to test the hypothesis that exposure to a care management intervention increased the likelihood of having the appropriate medication or procedures performed, as well as increased medication persistency. Statistically significant clinical metric improvements occurred in each of the 5 conditions studied. Increased medication persistency was found for beta-blocker medication for members with coronary artery disease, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker and diuretic medications for members with heart failure, bronchodilator and corticosteroid medications for members with chronic obstructive pulmonary disease, and aspirin/antiplatelet medications for members with diabetes. This study demonstrates that a care management program increases the likelihood of having an appropriate medication dispensed and/or an appropriate clinical test performed, as well as increased likelihood of medication persistency, in people with chronic conditions. (Population Health Management 2015;18:39–46)

Introduction

The Chronic Care Model (CCM) is a framework to guide organizational and quality changes in health care delivery to improve care for people with chronic conditions such as hypertension, congestive heart failure, depression, diabetes, or asthma. The CCM framework evolved in response to growing numbers of people with chronic conditions and the inadequacies of the existing health care system to treat those with chronic conditions. The existing health care system was designed to treat acute illness rapidly and efficiently, but failed to provide optimal care to people with chronic conditions or to assist people to become self-managers of their illnesses.¹

The CCM paradigm's basis for health care delivery is the member-provider partnership involving collaboration of care and self-management education and support.² This is in contrast to the existing health care delivery model's deficiencies in managing chronic conditions, which included rushed practitioners not following established practice guidelines, lack of care coordination or active follow-up to ensure the best outcomes, and patients inadequately trained to manage their own illnesses.³

Care management emerged as an application of the CCM in the mid-1990s and as a strategy to help mitigate these deficiencies using nurses as extra practice resources engaged in care coordination and teaching self-management skills.^4,5 Although a consensus definition of care management remains ambiguous,^4,6 care management programs do share the primary objectives of improving quality, consistency, and comprehensiveness of cost-effective care for people with chronic illnesses.

One strategy for care management programs to achieve better patient outcomes and to reduce costs is through better adherence to medication therapy regimens.⁷ Medication therapy is a major component of treating a variety of acute and chronic medical conditions, and the success of a medication in producing the desired outcomes depends on adhering to the prescribed regimen. Lack of adherence to the prescribed regimen has long been considered a roadblock to achieving better patient outcomes.⁸ The measurement of adhering to the prescribed regimen is captured by measuring medication persistency, which is the amount of time that an individual remains on a chronic drug therapy, whereby persistent individuals refill their medications regularly.⁹

In the attempt to achieve better patient outcomes and reduce costs, billions of dollars have been spent in the United States on the development and implementation of care management programs.¹⁰ As many as 38 states have engaged in care management programs for a portion of their Medicaid population¹¹ and many self-insured employers have offered care management programs to their employees, often with mixed results.¹²

The key contribution of this study is to estimate the effects of a population-based care management program across multiple comorbidities for persons with chronic conditions in a state Medicaid plan. Not only does this study add to the literature for state Medicaid care management evaluations but, in general, it adds to the literature on care management clinical metrics and medication persistency.

Methods

Subjects

On July 1, 2006, the state of Illinois implemented a care management program, Your Healthcare Plus. The target population included the adult aged, blind, and disabled Medicaid-only population living in the community (noninstitutionalized). In fiscal year 2007, this target population under care management represented 4.41% of the state's Medicaid program enrollees, but accounted for over 18% of the state's Medicaid expenditures, as calculated by the state's administrative claims analysis.

Subjects were selected from administrative claims diagnoses incurred between July 2005 and June 2009 for the aged, blind, and disabled Medicaid population living in noninstitutionalized settings. Additionally, subject inclusion criteria required having been diagnosed with one of the following 5 chronic conditions following the Population Health Alliance Guidelines v.5¹³: asthma, coronary artery disease (CAD), chronic obstructive pulmonary disease (COPD), diabetes, or congestive heart failure (CHF). The total number of subjects eligible for the intervention was 32,334. Of that number, 15,235 received the intervention at some point between July 2006 and June 2009.

Intervention

In July 2006, registered nurses (RNs) began calling identified members with one or more of the 5 chronic conditions (asthma, diabetes, CHF, CAD, or COPD) for program enrollment. Interventions focused on patient self-management and closing gaps in evidence-based care. Members were not randomized into an intervention or control group. For those members who chose to enroll, program services and interventions included the following: a member-centric and customized self-management plan of care, formal education sessions where nurses taught specific care management skills to members, and the availability of 24-hour access to nurse counseling and symptom advice. In addition, enrollees received individualized assessment summary letters and reminders (for medication compliance and vaccination). Addressing gaps in care started with obtaining the appropriate test/s and/or prescribing indicated medications. For optimal effects and outcomes, the patient must be persistent with the medication use and/or plan of testing. A multipronged and innovative provider engagement strategy was implemented. Physicians received clinical alerts about member medication reported usage and claims data summaries for their Medicaid members. This was coupled with information about medication and testing indicating gaps in national clinical guideline recommendations. The summary data on gaps in care identified from claims was updated quarterly.

Participating members received multiple nurse contacts throughout the course of the intervention. Examples of these are described in the following sections and include (1) staff-member assessments (initial, biannual, and annual), (2) staff-member monitoring/education calls, and (3) staff-member symptomatic calls. Members were supported to identify and work with a medical home primary care provider (PCP). As needed, staff arranged transportation and even accompanied members to provider appointments, not only for acute needs but, importantly, for preventive care.

Scheduled member assessments

The nursing staff-member assessments (initial, biannual, and annual) included gathering of member self-reported information on areas such as medication use; adherence and barriers; biometric data, such as current weight and recent tests; immunizations; knowledge of their chronic condition(s) and symptoms; current condition self-management practices; use of medical home; and recent emergency department (ED) or inpatient utilization. Assessments were always conducted by an RN. Although most assessments were conducted telephonically, some were conducted face-to-face in the member's home, the PCP's office, or during acute inpatient stays. Upon completion of the assessment, an individualized care plan was created. The plan of care was then shared with the member, the provider, and the rest of the multidisciplinary care team.

Scheduled member monitoring/education calls

The staff-member monitoring/education contacts were planned (scheduled) sessions that occurred in the interim period between assessments. Typical activities of this contact type included addressing the member's care plan problems and addressing any other pressing member concerns. Motivational interviewing techniques were used by staff to facilitate needed behavior changes for effectively self-managing the chronic condition(s).

Condition-specific education was provided. Depending on the member's needs, this education may have included explanations of the health condition(s), medication use, the benefits of persistence, and how to recognize symptoms of decompensation and take appropriate actions. Although the primary staff were RNs, monitoring calls, care coordination activities, and peer support functions were conducted by other multidisciplinary care team staff including social workers, community health workers, and behavioral health specialists.

Unscheduled member symptomatic calls

The staff-member symptomatic calls were ad hoc or unscheduled contacts that occurred in addition to the scheduled monitoring/education contacts already described. These were next-day follow-up contacts that occurred because a member reported actual symptoms or another urgent or concerning issue. Examples include follow-up with the member after a visit to a provider, or advice when a member wants to discuss current symptoms or other more urgent needs, medication side effects, need for refills, and transportation coordination.

Research design

This study is a quasi-experimental, nonrandomized experimental and control group design with identified independent and dependent variables of interest. Multivariate regression analysis was used to test the hypothesis that exposure to the care management program changed: (1) the likelihood of having an appropriate medication dispensed, (2) the likelihood of having an appropriate clinical test performed, and (3) the likelihood of increasing medication persistency for a prescribed medication that was initially filled.

For this analysis, the unit of measurement is the individual Medicaid member eligible on the last day of each study period. For each study period, administrative claims were used from the 12 months prior to the last day of each study period to determine compliance with clinical medication and procedure testing for each member.

The intervention started in July 2006. As such, for the first program year, people were selected based on Medicaid eligibility on June 30, 2007. For these members, administrative claims were used to determine medication use and clinical testing between July 1, 2006, and June 30, 2007, allowing for a 6-month claims runout.

Independent and dependent variable: clinical metric analysis

The dependent variable for each regression was the dichotomous indicator of whether or not the member had the specified medication prescription filled or clinical testing procedure performed.

The independent or explanatory variables included (1) a referral indicator showing whether or not a member had been referred to a particular care management program (asthma, diabetes, CHF, CAD, or COPD), (2) an exposure indicator showing whether or not a member had participated in a particular care management program, (3) a high-risk indicator (members identified as having cancer, end-stage renal disease, human immunodeficiency virus, hemophilia, traumatic brain injury, or a previous organ transplant), (4) age/sex grouping, (5) the predictive model acute index and chronic index scores prior to each member being referred into a care management program calculated by the predictive modeling company MEDai,¹⁴ and (6) a time variable.

The statistical measure used was the odds ratio as an approximation of the relative risk ratio.¹⁵ Subtracting the relative risk ratio from one is the preventive fraction in the exposed group (efficacy).¹⁵ This is the percent of subjects in the exposed group who meet the clinical metric who would not have satisfied the clinical metric without exposure. For example, an odds ratio of 1.23 for asthma controller medication indicates that among the exposed group (people exposed to the asthma care management program), 23% fewer would have received asthma controller medication in the absence of the care management program.

In addition to the odds ratio, the population attributable risk is calculated. The population attributable risk is the proportion of cases (people with the medication filled or procedure performed) in the entire population (ie, cases both exposed and unexposed to the intervention) that is attributable to the exposure to the care management program.¹⁵

Medication persistency analysis

Medication persistency is defined as the number of days a member remains on drug therapy. This can be shown in administrative claims by using the number of days supply of the drug. The Cox proportional hazard function¹⁶ was used to estimate the difference in medication persistency between each group (exposure to the care management program and control) for each medication class.

The dependent variable for each regression was the number of days a member was persistent in taking their medication. A gap of 30 days is allowed between the end of a prescription's days supply and the filling of the next prescription for the same medication. For example, if on May 1, a 30-day supply for an angiotensin-converting enzyme (ACE) inhibitor is filled for the first time in the study period, and on June 20 another 30-day supply is filled with no other refills, this person is persistent between May 1 and July 19 (June 20+30 days). Once a person has been classified as nonpersistent by exceeding the allowable refill gap, any subsequent refilling behavior is no longer considered.⁹ Censoring was taken into account by the model and occurred when a person continued to take the medication through the end of the study period. It is not known when the person eventually (if at all) became nonpersistent because the study period ended. The hazard function measures the time from the start of an event (the first prescription) to the end of the event (the person is no longer persistent).

The independent or explanatory variables include (1) an exposure indicator showing whether or not a member had participated in a particular care management program, (2) an indicator of sex, (3) age variable, (4) number of inpatient admissions, (5) number of ED visits, (6) Charlson comorbidity index,¹⁷ (7) total annual prescription drug costs, (8) total annual medical costs, and (9) the predictive model acute index score prior to each member being referred into a care management program calculated by the predictive modeling company MEDai.

For persistency, the statistical measure is the hazard ratio. The hazard here is ceasing to fill the prescribed medication. The hazard ratio is a measure of relative hazard between the group exposed to a care management program and the control group. For example, a hazard ratio of 1.0 indicates that the hazard for ceasing to fill the prescribed medication for the intervention group is equal to the hazard for ceasing to fill the medication for the control group. Any hazard ratio less than 1.0 indicates a lower hazard of ceasing to fill the prescribed medication for the intervention group. In other words, the care management group is more persistent in filling their prescribed drug therapy.

Results

Descriptive statistics

Over the 4 years of this intervention, 32,334 members were identified with one of the 5 chronic conditions and referred for the intervention. Within this intervention, 15,235 were managed by direct nurse contact at some point over the 4-year intervention time period. Table 1 shows descriptive statistics for the fourth year of the intervention.

Table 1.

Descriptive Statistics for Year 4

Condition	Metric	Population (n)	Clinical Metric Met (n, %)		Population Exposed to Intervention (n, %)
Asthma	Controller Medication	7015	5327	75.9	2385	34.0
	Corticosteroid Medication^*	4037	2520	62.4	1337	33.1
CAD	ACE Inhibitor/ARB Medication	1458	1028	70.5	691	47.4
	Aspirin/Antiplatelet Medication	1458	1082	74.2	691	47.4
	Beta-Blocker Medication^†	93	29	31.2	37	39.8
	Cholesterol Blood Test	1458	1151	78.9	691	47.4
	Pneumococcal Vaccination	1458	200	13.7	691	47.4
	Statin Medication	1458	1210	83.0	691	47.4
Heart Failure	ACE Inhibitor/ARB Medication	2593	2174	83.8	1522	58.7
	Aspirin/Antiplatelet Medication	2593	1701	65.6	1522	58.7
	Beta-Blocker Medication	2593	1999	77.1	1522	58.7
	Diuretic Medication	2593	2159	83.3	1522	58.7
	Pneumococcal Vaccination	2593	311	12.0	1522	58.7
COPD	Bronchodilator Medication^‡	549	505	92.0	382	69.6
	Corticosteroid Medication^§	1632	1375	84.3	1130	69.2
	Pneumococcal Vaccination	4141	694	16.8	2744	66.3
	Spirometry Test	4141	1786	43.1	2744	66.3
Diabetes	Cholesterol Blood Test	7006	5354	76.4	3174	45.3
	HbA1c Blood Test	7006	5761	82.2	3174	45.3
	Urine Protein Test	7006	4685	66.9	3174	45.3
	Retinal Eye Exam	7006	2822	40.3	3174	45.3
	Aspirin/Antiplatelet Medication	7006	3125	44.6	3174	45.3
	Statin Medication	7006	4739	67.6	3174	45.3

ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease.

For members with acute asthma exacerbations, defined as an inpatient admission/emergency department visit/3 or more dispensed prescriptions for short-acting beta-agonist within a 3-month period.

†

For members with a prescription 30 days after a myocardial infarction.

‡

For members with a COPD inpatient admission.

For members with a COPD exacerbation defined as diagnosis code 491.21.

For each of the 22 clinical metrics, listed are:

1. Number in the population

2. Number and percent of people meeting that particular clinical metric (ie, members who had the presence of appropriate medication dispensed or clinical test performed)

3. Number and percent of people exposed to the intervention (ie, members managed through direct nurse contact)

The percent of people meeting the particular clinical metrics ranged from a low of 13.7% for the CAD pneumococcal vaccination metric to a high of 92.0% for the COPD bronchodilator medication metric. Furthermore, exposure to the intervention ranged from a low of 33.1% for the asthma corticosteroid medication population to a high of 69.6% for the COPD bronchodilator medication population.

Multivariate regression results

• Table 2 shows the odds ratios and levels of significance for each clinical metric and condition.

• Table 3 shows the population attributable risk for those metrics with a statistically significant odds ratio.

• Table 4 shows the hazard ratio for the members who have taken at least one of the medications in the drug class.

• Supplementary figures 1–14 show the survival curves from the Cox regression stratified by care management exposure. The supplementary figures are available in the online article at www.liebertpub.com/pop.

Table 2.

Odds Ratios by Clinical Metric by Year

Condition	Metric	Year 1 Odds Ratio	Year 2 Odds Ratio	Year 3 Odds Ratio	Year 4 Odds Ratio
Asthma	Controller Medication	1.38^***	1.36^***	1.38^***	1.23^***
	Corticosteroid Medication	0.99	1.16^*	0.99	0.97
CAD	ACE Inhibitor/ARB Medication	1.24	0.97	1.09	0.97
	Aspirin/Antiplatelet Medication	1.35^**	1.26^*	1.29^**	1.03
	Beta-Blocker Medication	0.80	2.44^*	0.88	0.75
	Cholesterol Blood Test	1.38^**	1.61^***	1.60^***	1.25^*
	Pneumococcal Vaccination	2.12^***	2.09^***	1.67^***	1.86^***
	Statin Medication	1.51^***	1.53^***	1.54^***	1.42^**
Heart Failure	ACE Inhibitor/ARB Medication	1.47^***	1.85^***	1.34^**	1.56^***
	Aspirin/Antiplatelet Medication	1.26^**	1.33^***	1.31^***	1.08
	Beta-Blocker Medication	1.42^***	1.59^***	1.32^***	1.32^***
	Diuretic Medication	1.48^***	1.41^***	1.33^**	1.61^***
	Pneumococcal Vaccination	1.51^**	1.62^***	1.87^***	1.81^***
COPD	Bronchodilator Medication	5.12^***	2.52^**	3.72^**	5.05^***
	Corticosteroid Medication	2.82^***	1.78^***	2.48^***	1.64^***
	Pneumococcal Vaccination	1.93^***	2.47^***	2.03^***	1.87^***
	Spirometry Test	2.28^***	1.93^***	1.66^***	1.63^***
Diabetes	Cholesterol Blood Test	1.26^***	1.22^***	1.31^***	1.32^***
	HbA1c Blood Test	1.51^***	1.5^***	1.37^***	1.43^***
	Urine Protein Test	1.03	1.07	0.94	1.04
	Retinal Eye Exam	1.28^***	1.36^***	1.25^***	1.39^***
	Aspirin/Antiplatelet Medication	0.94	0.90^**	0.91^*	0.83^***
	Statin Medication	1.06	1.20^***	1.13^**	1.12^**

ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease.

P value≤0.10.

P value≤0.05.

***

P value≤0.01.

Table 3.

Population Attributable Risk (PAR) by Year for Statistically Significant Metrics

Condition	Metric	Year 1 PAR (%)	Year 2 PAR (%)	Year 3 PAR (%)	Year 4 PAR (%)
Asthma	Controller Medication	7.4	8.6	10.5	7.3
	Corticosteroid Medication		3.9
CAD	ACE Inhibitor/ARB Medication
	Aspirin/Antiplatelet Medication	10.0	8.7	11.0
	Beta-Blocker Medication		35.2
	Cholesterol Blood Test	10.8	18.1	20.5	10.5
	Pneumococcal Vaccination	26.3	28.3	22.3	29.1
	Statin Medication	13.9	16.2	18.9	16.6
Heart Failure	ACE Inhibitor/ARB Medication	15.1	27.6	14.6	24.8
	Aspirin/Antiplatelet Medication	9.0	12.9	13.4
	Beta-Blocker Medication	13.6	21.0	13.8	15.7
	Diuretic Medication	15.3	15.5	14.3	26.3
	Pneumococcal Vaccination	16.2	21.7	30.3	32.2
COPD	Bronchodilator Medication	72.9	49.6	63.1	73.8
	Corticosteroid Medication	53.3	32.2	48.6	30.8
	Pneumococcal Vaccination	33.6	45.9	38.3	36.6
	Spirometry Test	41.1	35.0	28.5	29.3
Diabetes	Cholesterol Blood Test	8.0	7.7	11.3	12.7
	HbA1c Blood Test	14.5	16.2	13.3	16.4
	Urine Protein Test
	Retinal Eye Exam	8.6	12.3	9.5	15.1
	Aspirin/Antiplatelet Medication		−4.2	−3.9	−8.3
	Statin Medication		7.3	5.1	5.0

ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease.

Table 4.

Medication Persistency Hazard Ratios for Year 4

Condition	Metric	Population with Rx	Percent Managed	Hazard Ratio
Asthma	Controller Medication	5327	35%	0.938^*
	Corticosteroid Medication	2520	33%	0.978
CAD	ACE Inhibitor/ARB Medication	1028	47%	0.898
	Aspirin/Antiplatelet Medication	1082	48%	0.988
	Beta-Blocker Medication	29	34%	0.052^***
	Statin Medication	1210	49%	0.939
Heart Failure	ACE Inhibitor/ARB Medication	2174	60%	0.814^***
	Aspirin/Antiplatelet Medication	1701	60%	0.916
	Beta-Blocker Medication	1999	60%	0.892^*
	Diuretic Medication	2159	61%	0.853^***
COPD	Bronchodilator Medication	505	72%	0.747^**
	Corticosteroid Medication	1375	71%	0.828^***
Diabetes	Aspirin/Antiplatelet Medication	3125	43%	0.892^**
	Statin Medication	4739	46%	0.887^***

ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; Rx, prescription.

P value≤0.10.

P value≤0.05.

***

P value≤0.01.

It is interesting to note that persistency falls at 30 days, indicating that many people do not refill their 30-day prescription after the initial fill. For example, the percent of CAD non-managed members becoming nonpersistent after 30 days for the ACE inhibitor/angiotensin receptor blocker (ARB) medication is 12.3%. Similarly, for heart failure non-managed members, 18.1% become nonpersistent for the same medication after 30 days.

Asthma results

The 2 asthma clinical metrics are as follows: controller medication for all asthmatics, and corticosteroid medication for members with acute asthma exacerbations. Exacerbations were an asthma inpatient or ED visit or 3 or more dispensed prescriptions for short-acting beta-agonist within a 3-month period.

The odds ratio was above 1 and statistically significant at the 1% level for the controller medication for all 4 years (Table 2). Of note, without the care management intervention, 7.3% fewer members would have had a prescription for controller medication (Table 3).

The odds ratio for the corticosteroid metric was near 1 for all 4 years and not statistically significant at the 5% level for any year. Therefore, the intervention had no statistically measurable impact for this clinical metric. One limitation specific to this metric is that hospital-dispensed medications and samples, which are given to patients in a hospital setting during exacerbations, are not included.

Of the 2 medication categories, only controller medication shows a statistically significant hazard ratio indicating that the intervention group had a higher medication persistency (Table 4). The corticosteroid medication to treat acute asthma exacerbations is time limited by design and not expected to have a high persistency.

CAD results

The 6 clinical metrics are as follows: ACE inhibitor/ARB medication, aspirin/antiplatelet medication, beta-blocker medication for members with a prescription 30 days after a myocardial infarction, cholesterol blood test, pneumococcal vaccination, and statin medication.

Pneumococcal vaccination and the statin medication metrics were statistically significant at at least the 5% level for all 4 years, indicating that the intervention had a statistically measurable impact (Table 2). Table 3 shows that without the care management intervention, 16.6% fewer members would have had a prescription for statin medication.

For the 4 medication categories, only beta-blocker medication shows a statistically significant hazard ratio indicating that the intervention group had a higher medication persistency (Table 4).

Heart failure results

The 5 clinical metrics are as follows: ACE inhibitor/ARB medication, aspirin/antiplatelet medication, beta-blocker medication, diuretic medication, and pneumococcal vaccination.

All clinical metrics were statistically significant at the 5% level for all years except for the aspirin/antiplatelet medication metric in year 4 (Table 2). Table 3 shows that without the care management intervention, 26.3% fewer members would have had a prescription for diuretic medication.

For the 4 medication categories, 3 show a statistically significant hazard ratio: ACE inhibitor/ARB, beta-blocker, and diuretic medication (Table 4). This shows that the intervention group had a higher medication persistency.

COPD results

The 4 clinical metrics are: bronchodilator medication for members with a COPD inpatient admission, corticosteroid medication for members with a COPD exacerbation (defined as diagnosis code 491.21), pneumococcal vaccination, and spirometry test.

All clinical metrics were statistically significant at the 5% level for all years (Table 2). Of note, without the care management intervention, 29.3% fewer members would have had a a spirometry test (Table 3).

Both of the medication categories of bronchodilator and corticosteroid medication have a statistically significant hazard ratio, showing that the intervention group had a higher medication persistency (Table 4).

Diabetes results

The 5 clinical metrics are as follows: cholesterol blood test, HbA1c blood test, urine microalbumin test, retinal eye exam, aspirin/antiplatelet medication, and statin medication.

Three clinical metrics were statistically significant at the 5% level for all years. The diabetes statin medication clinical metric was statistically significant for years 2 through 4 (Table 2). It is noteworthy that without the care management intervention, 5.0% fewer members would have had a prescription for statin medication (Table 3).

Both of the medication categories of aspirin/antiplatelet medication and statin medication have a statistically significant hazard, showing that the intervention group had a higher medication persistency (Table 4).

Discussion

Findings

This study supports statistically significant associations of exposure to the care management program and increased medication use and testing procedures performed for members diagnosed with asthma, CAD, CHF, COPD, or diabetes. This study also supports care management programs increasing the likelihood of having an appropriate medication dispensed or clinical test performed, and increasing the likelihood of medication persistency.

Prior work

Previous studies specific to state Medicaid programs include those in Florida,^18,19 Indiana,^20,21 New York City,²² Washington,²³ and Tennessee.²⁴ With the exception of the Tennessee study, these did not examine the impacts on clinical metrics (ie, the presence of appropriate medication dispensed or clinical test performed) or medication persistency (ie, the length of time a person takes a drug without ceasing). The Tennessee study evaluated clinical metrics specific to people with diabetes showing an improvement. Other community-based Medicaid studies have shown mixed clinical metric results but have not shown medication persistency.^25,26,27 One Medicaid study evaluated medication persistency to the extent of determining the factors associated with persistency rather than an evaluation of a care management program.²⁸

Strengths

A 2-pronged approach is necessary to have a positive impact on health outcomes because of the diverse nature of human behavior complicated by complex situations and chronic health issues. This approach must focus on (1) supporting and motivating the individual member, and (2) supporting and motivating a health care delivery system that has been remunerated at lower levels for caring for this population in lieu of other populations with private insurance and/or Medicare.

Therefore, successful care management programs must work with providers to support improvements in the quality of care delivered. Because medications must first be prescribed by a provider and clinical tests performed by providers, the growing evidence supporting a medical home may be a necessary component to effective care management,^5,29 albeit with some mixed results.³⁰ Because care management programs facilitate better adherence to evidence-based clinical guidelines and promote coordinated efforts and communication avenues between the multiple providers invested in the member's care,^1,2 there will be increased savings and quality of care as demonstrated in this study. The strength of a 4-year study duration coupled with the large population and sample size allowed for statistically significant results.

Limitations

The chief challenge was the ability to maintain contact over time; this was largely because of population mobility and eligibility churn. Every effort was made to maintain contact including using community health workers to locate members, gathering updated member demographic data from provider offices/clinics, and obtaining demographic data from a hospitalization event through utilization management sources.

This study's limitation is its lack of ability to be generalized, as is true with all quasi-experimental designs and limited control studies. Multivariate regression analysis was used to increase control for potential confounding variables because randomization was not possible. Additionally, this study was specific to a state Medicaid population. Translating this to another Medicaid population or a commercially insured population has yet to be determined; however, given the success of care management in both commercial and Medicaid populations, there is reason to be optimistic that the results are relevant within these populations.

Suggestions for future research

Expansion of future research to another Medicaid population or a commercially insured population would improve generalization. Because this was a high-intensity care management program, the results in a less intense care management intervention are unknown, as is whether or not favorable clinical results can be maintained.

This study supports care management programs increasing the likelihood of the appropriate medication being dispensed, an appropriate clinical test performed, and increased medication persistency. Improving clinical metrics and increasing medication persistency by modalities such as care management should be studied in other populations and settings in order to link improved clinical testing and medication persistency with improved health outcomes, increased functionality, lower cost, and lower hospital usage.

Footnotes

Author Disclosure Statement

Drs. Berg, Leary, and Medina, Mr. Donnelley, and Ms. Warnick declared the following conflicts of interest with respect to the research, authorship, and/or publication of this article: Drs. Berg, Leary, Mr. Donnelly, and Ms. Warnick are employed by McKesson Corporation, which administered the intervention for the State of Illinois. Dr. Medina is employed by the State of Illinois. The authors received no financial support for the research, authorship, and/or publication of this article.

References

Wagner

, Austin

, Davis

, Hindmarsh

, Schaefer

, Bonomi

. Improving chronic illness care: translating evidence into action. Health Aff (Millwood)., 2001; 20(6):64–78.

Bodenheimer

, Lorig

, Holman

, Grumbach

. Patient self-management of chronic disease in primary care. JAMA., 2002; 288:2469–2475.

Improving Chronic Illness Care. The chronic care model. http://www.improvingchroniccare.org/index.php?p=The_Chronic_Care_Model&s=2. Accessed September 16, 2013 .

Krumholz

, Curie

, Riegel

, et al. A taxonomy for disease management: A scientific statement from the American Heart Association disease management taxonomy writing group. Circulation., 2006; 114:1432–1445.

Rosenthal

. The medical home: growing evidence to support a new approach to primary care. J Am Board Fam Med., 2008; 21:427–440.

Population Health Alliance. Understanding population health. http://www.populationhealthalliance.org/research/understanding-population-health.html. Accessed April 8, 2014 .

Jha

, Aubert

, Yao

, Teagarden

, Epstein

. Greater adherence to diabetes drugs in linked to less hospital use and could save nearly $5 billion annually. Health Aff (Millwood)., 2012; 31:1836–1846.

Cutler

, Everett

. Thinking outside the pillbox—mediation adherence as a priority for health care reform. N Engl J Med., 2010; 362:1553–1555.

Sikka

, Xia

, Aubert

. Estimating medication persistency using administrative claims data. Am J Manag Care., 2005; 11:449–457.

10.

Mattke

, Seid

, Ma

. Evidence for the effect of disease management: is $1 billion a year a good investment?. Am J Manag Care., 2007; 13:670–676.

11.

National Conference of State Legislatures. State Medicaid disease management program descriptions. http://www.ncsl.org/IssuesResearch/Health/StateMedicaidDiseaseManagementPrograms/tabid/14421/Default.aspx. Accessed March 25, 2011 .

12.

Holtz-Eakin

. An analysis of the literature on disease management programs. http://www.cbo.gov/publication/16076. Accessed April 7, 2014 .

13.

Population Health Alliance. Outcome Guidelines Report v.5. https://netforum.avectra.com/eWeb/DynamicPage.aspx?Site=DMAA&WebCode=DMAAPublications. Accessed April 8, 2014 .

14.

MEDai. Risk Navigator Clinical. http://www.medai.com/products/risk-navigator-solutions/risk-navigator-clinical/. Accessed April 8, 2014 .

15.

Gregg

. Field Epidemiology. 1st ed. New York: Oxford University Press; 1996:103–106.

16.

Cox

. Regression models and life-tables. J Royal Statistical Society. Series B., 1972; 34:187–220.

17.

Charlson

, Pompei

, Ales

, MacKenzie

. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis., 1987; 40:373–383.

18.

Afifi

, Morisky

, Kominski

, Kotlerman

. Impact of disease management on health care utilization: Evidence from the “Florida: A Health State (FAHS)” Medicaid program. Prev Med., 2007; 44:547–553.

19.

Kominski

, Morisky

, Afifi

, Kotlerman

. The effect of disease management on utilization of services by race/ethnicity: evidence from the Florida Medicaid program. Am J Manag Care., 2008; 14:168–172.

20.

Holmes

, Ackermann

, Zillich

, Katz

, Downs

, Inui

. The net fiscal impact of a chronic disease management program: Indiana Medicaid. Health Aff (Millwood)., 2008; 27:855–864.

21.

Katz

, Holmes

, Stump

, et al. The Indiana Chronic Disease Management Program's impact on Medicaid claims: a longitudinal, statewide evaluation. Med Care., 2009; 47:154–160.

22.

Billings

, Mijanovich

. Improving the management of care for high-cost Medicaid patients. Health Aff (Millwood)., 2007; 26:1643–1655.

23.

Lind

, Kaplan

. An evaluation of Washington's Medicaid disease-management program. Manag Care Interface., 2007; 20(10):45–50.

24.

Patric

, Stickles

, Turpin

, et al. Diabetes disease management in Medicaid managed care: A program evaluation. Dis Manag., 2006; 9:144–156.

25.

Berg

, Wadhwa

. Diabetes disease management results in Hispanic Medicaid patients. J Health Care Poor Underserved., 2009; 20:432–443.

26.

Anderson

, Christison-Lagay

, Villagra

, Liu

, Dzuura

. Managing the space between visits: a randomized trial of disease management for diabetes in a community health center. J Gen Intern Med., 2010; 25:1116–1122.

27.

Johnson

, Berg

, Long

, Wadhwa

. A matched-cohort study of utilization outcomes for an adult Medicaid population enrolled in an asthma disease management program. J Ambul Care Manage., 2007; 30:241–58.

28.

Khanna

, Pace

, Mahabaleshwarkar

, Basak

, Datar

, Banahan

. Medication adherence among recipients with chronic diseases enrolled in a state Medicaid program. Popul Health Manag., 2012; 15:253–260.

29.

Rosenthal

, Friedberg

, Singer

, Eastman

, Li

, Schneider

. Effect of a multipayer patient-centered medical home on health care utilization and quality: the Rhode Island chronic care sustainability initiative pilot program. JAMA Intern Med., 2013; 173:1907–1913.

30.

Friedberg

, Schneider

, Rosenthal

, Volpp

, Werner

. Association between participation in a multipayer medical home intervention and changes in quality, utilization, and costs of care. JAMA., 2014; 311:815–825.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.55 MB