Improving the Care of Dual Eligible Patients in Rural Federally Qualified Health Centers

Abstract

Background: Dual eligible persons are those covered by both Medicare and Medicaid. There were 9.6 million dual eligible persons in the United States and 82 000 in West Virginia in 2010. Dual eligibles are poorer, sicker, and more burdened with serious mental health conditions than Medicare or Medicaid patients as a whole. Their health care costs are significantly higher and they are more likely to receive fragmented ineffective care. Purpose: To improve the care experience and health care outcomes of dual eligible patients by the expanded use of care coordinators and clinical pharmacists. Methods: During 2012, 3 rural federally qualified community health centers in West Virginia identified 200 dual eligible patients each. Those with hospitalizations received more frequent care coordinator contacts. Those on more than 15 chronic medications had drug utilization reviews with recommendations to primary care providers. Baseline measures included demographics, chronic diseases, total medications and Beers list medications, hospitalization, and emergency room (ER) use in the previous year. Postintervention measures included hospitalization, ER use, total medications, and Beers list medications. Results: Out of 556 identified patients, 502 were contacted and enrolled. Sixty-five percent were female. The median age was 69 years, with a range of 29 to 93 years. Nineteen percent (19%) of patients were on 15 or more medications, 56% on psychotropic medication, and 33% on chronic opiates. One site showed reductions of 34% in hospitalizations and 25% in ER visits during the intervention year. For all sites combined, there was a 5.5% reduction in total medications and a 14.8% reduction in Beers list medications. Conclusions: A modest investment in care coordination and clinical pharmacy review can produce significant reductions in hospitalization and harmful polypharmacy for community dwelling dual eligible patients.

Introduction

There are 9.6 million dual eligible persons in the United States; 82 000 in West Virginia. ¹ These patients are disabled, poorer, sicker, more likely to be disabled, and more burdened with serious mental health conditions than Medicare or Medicaid patients as a whole.^1,2 They consume 1.7 their per capita share of Medicare and 2.6 of Medicaid resources.^3-5 The total cost of their care was $319.5 billion in 2011. ² Medical care for dual eligible patients is often substandard and fragmented. This is partly due to the conflicting rules and incentives of the Medicare and Medicaid programs each of which pays for different parts of dual eligibles’ care. ²

A network of federally qualified health centers (FQHCs) and a tertiary care referral hospital in southern West Virginia undertook a care coordination project for a sample of their dual eligible patients. The partners had more than 10 years of experience working together to improve the care of high-risk populations including those with no insurance, frail elder status, disability, chronic obstructive pulmonary disease, and black lung.

Methods

Setting

The partners in this project included 3 participating FQHCs with a combined 50 000 users and 190 000 patient encounters in 2011. The tertiary care referral hospital had 800 beds with more than 40 000 inpatient discharges and 560 000 outpatient visits in 2011. These partners held a series of facilitated meetings during 2011 to develop a plan for coordinating the care of high-risk populations with multiple chronic conditions. A decision was made to focus on the dual eligible population. A care model was developed based on the strategy of “a triple arrow for the triple aim”. ⁶ The “triple arrow” was “relationships, transitions, medications.”

Patient Selection

Each FQHC was asked to recruit 200 dual eligible patients. This was a purposeful convenience sample. Lists were generated by identifying eligible patients in a serial manner until the number 200 was reached. Site 3 only identified 156 patients.

The Care Model

Relationships

The “relationships” part of this model posited that a close relationship with extra communication and advocacy would better serve the needs of high-risk patients. The use of care coordinators was already established, to varying degrees, in the 3 participating FQHCs. Each practice assigned at least 0.5 FTE (full-time equivalent) care coordinator to its dual eligible patients selected for participation. The care coordinator conducted a structured review of the medical record, an in-person “welcome interview”, and regular telephone contact on an as-needed basis with each patient throughout the 2012 intervention period. The care coordinator brokered optimal access and 2-way communication between the patients and the primary care team in the FQHC.

Transitions

Care coordinators were to review daily notifications of hospital admissions and emergency room (ER) visits and to contact the patient within 2 working days. They discussed discharge medications, follow-up appointments, and answered questions. They communicated unmet needs to the primary care team and made referrals to community resources.

Medications

The prevalence of polypharmacy ⁷ and the use of potentially inappropriate medications^8-10 are well documented in US chronic disease and elderly populations. Clinical pharmacists conducted drug utilization reviews (DURs) starting with patients on 15 or more medications. Criteria used in this DUR process included the validated STOPP and START criteria ⁷ and, for patients older than 65 years, the Beers list of potentially harmful medications. ⁸ The completed DURs were provided to the patient’s primary care provider with follow-up discussions and assistance to make indicated changes in medication regimens.

Data and Analysis

A set of baseline and outcome measures along with scannable data collection instruments were developed as part of this quality improvement project. The data forms completed by the care coordinators were collected and scanned to a database at the research institute of the main referral hospital. Hospitalization and ER utilization data were obtained from the referral hospital admissions data base with comparisons of the intervention year 2012 to the previous year of 2011. Medication lists were audited before and 3 to 4 months after the clinical pharmacists’ DUR.

All data were cleaned prior to performing analysis. Descriptive statistics, t test, chi-square, Fisher’s exact test, and McNemar test were used where deemed appropriate in presentation of data in the exhibits. A value is considered significant with a P value of <.05.

This study was approved by the institutional review board of Charleston Area Medical Center, Charleston, West Virginia.

Results

Patients

After medical record review, 529 patients were deemed eligible for participation. Sixty-five percent (65%) were female. The median age was 69 years with a range of 23 to 93 years. The prevalence of chronic diseases ranged from 73% for hypertension and 30% for diabetes to 4% for dementia. Fifty-six percent were diagnosed and treated for behavioral health disorders. Fifty-six percent were on 9 or more medications and 19% on 15 or more. Thirty-three percent were on chronic opiates.

Of this eligible cohort of 529 patients, 502 were actually contacted and enrolled.

Hospital and Emergency Room Utilization

Only site 1 had a sufficiently robust medical record exchange to obtain full data on pre- and postintervention admissions and ER visits. This was because the other 2 sites did not have sufficiently developed relationships with their referral hospitals to obtain matched data from the hospitals admissions databases. The data for this site are shown in Table 1.

Table 1.

Hospital and Emergency Room (ER) Use Pre and During Intervention: Site 1.

	2011		2012
	n	%	n	%	P
Total patients	193	100.0	183	100.0
1 or more hospital admits	30	15.5	23	12.6	<.001*
1 or more ER visits	40	20.7	30	16.4	<.001*
Total hospital admits	89	46.1	56	30.6	<.001*
Total ER visits	114	59.1	81	44.3	<.004*

*

Statistically significant (P < .05).

There were relative risk reductions of 34% in total hospital admissions and 25% in total emergency room visits, both significant.

Medication usage

Clinical pharmacists at all 3 sites were able to complete DURs on all patients who were prescribed 15 or more chronic medications. Three months later, a repeat medication list audit was completed to measure changes in total medications and Beers list medications. These results are shown in Table 2.

Table 2.

Medications Pre- and Postintervention.

	Site
	1	2	3	Totals
Patients on 15 or more medications	60	21	22	103
DURs completed	71	17	52	140
Total medications
Total medications pre	1096	243	686	2025
Total medications post	1019	203	692	1914
% change				−5.5, P = .001*
No. of DURs 65 and over	32	17	32	81
Beers medications pre	46	21	55	122
Beers medications post	38	14	52	104
% change				−14.8, P = .009*

Abbreviations: DUR. Drug utilization review.

*

Statistically significant (P < .05).

Total medications decreased at 2 sites and increased slightly at the third. Overall, there was a 5.5% decrease in total medications, P = .001. Beers list medications decreased at all 3 sites with an overall decrease of 14.8%, P = .009.

Discussion

This collaborative quality improvement project met a number of important process goals. A total of 502 patients had at least 1 contact with a personal care coordinator. Agreements with referral hospitals were established and timely notification of admissions and ER visits were initiated at all three sites. Patients, providers, and nurses gained greater awareness of the risks of polypharmacy and the role of clinical pharmacists to reduce it.

Potentially harmful medications were reduced in the older than 65 years group.

At 1 site, there were significant reductions in hospital and ER use.

Following the classification proposed by Rohrer, ¹¹ this would be considered a class D quality improvement study in which baseline conditions are identified, an intervention is undertaken, and monitoring is carried out to measure impact. The variable success in full implementation of the care model during this project spotlights the difficulties of carrying out quality improvement projects like this one in community-based settings.

Results from the Medicare Care Coordination Demonstration Project over the decade 2000 to 2010 showed that hospitalization could be reduced for targeted groups of Medicare beneficiaries in a cost neutral way. ¹² Brown et al ¹³ identified 6 features of successful programs. Our project used their lessons as its starting point.

This study had several limitations. First, the project used a convenience sample of dual eligible patients from each organization. Second, only 1 of the 3 sites was able to validly measure impact on hospital and ER utilization. Third, 3 months may have been too soon to perform post-DUR medication audits. Many patients would not yet have had a regular appointment when the primary care provider could make changes. Fourth, this project was carried out amid the full range of daily activities in busy primary care centers. Care coordinators were sometimes pulled away from scheduled activities with their dual-eligible panels to fill emergency staffing gaps. Fifth, patient satisfaction was not measured as part of this evaluation. Given the triple aim emphasis on “improving the experience of care,” patient satisfaction should be measured in future similar studies.

In conclusion, this quality improvement project suggests that a modest investment in care coordination and clinical pharmacy review can produce significant reductions in hospitalization and harmful polypharmacy for community-dwelling dual eligible patients.