Hawaii Medical Service Association Research Articles

Accuracy of Dermatologist Listings in Hawai'i's Medicaid (Med-QUEST) Physician Directories.

Since 2017, Hawai'i has had a statute requiring health plans to update their provider directories at least monthly. However, the results of this study suggest that despite this regulation, errors in physician directories may be an ongoing problem. Using publicly available online Medicaid physician directories from Med-QUEST, Hawaii Medical Service Association (HMSA), AlohaCare, 'Ohana Health Plan, and United Healthcare, 473 unique listings for dermatologists were reviewed and 411 (86.9%) of these listings contained at least 1 inaccuracy. Using the deficiency scoring methodology designed by the Centers for Medicare & Medicaid Services (CMS), it was found that the proportions of deficient listings were significantly different among the directories (P<.001). Med-QUEST had the highest weighted final deficiency score of 92.9% and HMSA had the lowest weighted final score of 49.2%. In between were United Healthcare (71.0%), 'Ohana Health Plan (69.7%), and AlohaCare (65.7%). It is unknown whether these results are an improvement from the implementation of the statute. Nevertheless, this issue can cause additional barriers for Medicaid patients who already experience narrower networks and longer wait times for dermatologists. Furthermore, it would also be worth investigating if this issue is also prevalent in listings for other specialties.

Designing a commercial bundle for cardiac procedures: The Percutaneous Coronary Intervention Episode Payment Model

BackgroundCardiac interventions account for a significant share of overall healthcare spending and have been the focus of several large-scale interventions to develop effective bundled payments. To date, however, none have proven successful in commercially insured populations. In 2018, we worked with Hawaii Medical Service Association (HMSA), the Blue Cross Blue Shield of Hawaii, to design a novel commercial bundled payment for percutaneous coronary interventions, the Percutaneous Coronary Intervention Episode Payment Model (PCI EPM). MethodsDescriptive analysis of HMSA's PCI EPM, including its inclusion criteria, contents of the bundle, target prices, shared savings model, and incentivized quality metrics. We also compare HMSA's PCI EPM to Medicare's Bundled Payment for Care Improvement programs and the cancelled Cardiac Care Model. ResultsHMSA's PCI EPM was designed through an iterative process with cardiologists and is the first commercial bundle to specifically target a cardiac procedure. PCI EPM incorporates site neutrality and incentivizes providers to shift care to the outpatient setting when medically permissible. Compared to existing non-commercial models, PCI EPM incorporate first-dollar shared savings and incentivized fewer quality metrics. ConclusionsReviewing features of the Percutaneous Coronary Intervention Episode Payment Model in comparison to existing Medicare programs is intended to help guide health plan and health policymakers when designing programs and policies related to cardiac interventions. ImplicationsBundled commercial payments for interventional cardiology procedures are promising and should continue to be further explored. Level of evidenceVI.

Designing a commercial medical bundle for cancer care: Hawaii Medical Service Association's Cancer Episode Model

BackgroundOncology care is expensive and exhibits substantial variation in cost and quality across clinicians and patients. Unlike many conditions with established bundled payment programs, cancer care includes a mix of inpatient and outpatient care that precludes hospital-based designs. In 2018, we worked with Hawaii Medical Service Association (HMSA), the Blue Cross Blue Shield of Hawaii, to design a novel commercial bundle for cancer care, the Cancer Episode Model. MethodsDescriptive analysis of HMSA's Cancer Episode Model, including its inclusion criteria, episode definitions, suite of enhanced services, shared savings model, and incentivized quality metrics. We also compare HMSA's Cancer Episode Model to Medicare's Oncology Care Model and three major commercial oncologic alternative payment models offered by Anthem, UnitedHealthcare, and Aetna. ResultsHMSA's Cancer Episode Model builds upon the successes and limitations of Medicare's Oncology Care Model and existing commercial alternative payment models. Compared to Medicare's Oncology Care Model, HMSA's Cancer Episode Model has stricter inclusion criteria, fewer incentivized quality metrics, a higher proportion of regional pricing, a different risk-adjustment model, and first-dollar shared savings. Compared to the majority of existing commercial models, HMSA's Cancer Episode Model includes total cost of care and a different risk-adjustment model. ConclusionsReviewing features of the Cancer Episode Model in comparison to other programs is intended to provide guidance to health plans and health policymakers in the design of programs and policies aimed at improving cancer care value. Level of evidenceLevel IV.

Changes in Outpatient Imaging Utilization and Spending Under a New Population-Based Primary Care Payment Model

ObjectiveTo evaluate whether the implementation of a new population-based primary care payment system, Population-Based Payments for Primary Care (3PC), initiated by Hawaii Medical Service Association (HMSA; the Blue Cross Blue Shield of Hawaii), was associated with changes in spending and utilization for outpatient imaging in its first year. MethodsIn this observational study, we used claims data from January 1, 2012, to December 31, 2016. We used a propensity-weighted difference-in-differences design to compare 70,284 HMSA patients in Hawaii attributed to 107 primary care physicians (PCPs) and 4 physician organizations participating in 3PC in its first year of implementation (2016) and 195,902 patients attributed to 312 PCPs and 14 physician organizations that used a fee-for-service model during the study period. The primary outcome was total spending on outpatient imaging tests, and secondary outcomes included spending and utilization by modality. ResultsThe study included 266,186 HMSA patients (mean age of 43.3 years; 51.7% women) and 419 PCPs (mean age of 54.9 years; 34.8% women). The 3PC system was not significantly associated with changes in total spending for outpatient imaging. Of 12 secondary outcomes, only 3 were statistically significant, including changes in nuclear medicine spending (adjusted differential change = −20.1% [95% confidence interval = −27.5% to −12.1%]; P < .001) and utilization (adjusted differential change = −18.1% [95% confidence interval = −23.8 to −11.9%]; P < .001). DiscussionThe HMSA 3PC system was not associated with significant changes in total spending for outpatient imaging, though spending and utilization on nuclear medicine tests decreased.

Association Between the Implementation of a Population-Based Primary Care Payment System and Achievement on Quality Measures in Hawaii

Hawaii Medical Service Association (HMSA), the Blue Cross Blue Shield of Hawaii, introduced Population-based Payments for Primary Care (3PC), a new capitation-based primary care payment system, in 2016. The effect of this system on quality measures has not been evaluated. To evaluate whether the 3PC system was associated with changes in quality, utilization, or spending in its first year. Observational study using HMSA claims and clinical registry data from January 1, 2012, to December 31, 2016, and a propensity-weighted difference-in-differences method to compare 77 225 HMSA members in Hawaii attributed to 107 primary care physicians (PCPs) and 4 physician organizations participating in the first wave of the 3PC and 222 233 members attributed to 312 PCPs and 14 physician organizations that continued in a fee-for-service model in 2016 but had 3PC start dates thereafter. Participation in the 3PC system. The primary outcome was the change in a composite measure score reflecting the probability that a member achieved an eligible measure out of 13 pooled Healthcare Effectiveness Data and Information Set quality measures. Primary care visits and total cost of care were among 15 secondary outcomes. In total, the study included 299 458 HMSA members (mean age, 42.1 years; 51.5% women) and 419 primary care physicians (mean age, 54.9 years; 34.8% women). The risk-standardized composite measure scores for 2012 to 2016 changed from 75.1% to 86.6% (+11.5 percentage points) in the 3PC group and 74.3% to 83.5% (+9.2 percentage points) in the non-3PC group (differential change, 2.3 percentage points [95% CI, 2.1 to 2.6 percentage points]; P < .001). Of 15 prespecified secondary end points for utilization and spending, 11 showed no significant difference. Compared with the non-3PC group, the 3PC system was associated with a significant reduction in the mean number of primary care visits (3.3 to 3.0 visits vs 3.3 to 3.1 visits; adjusted differential change, -3.9 percentage points [95% CI, -4.6 to -3.2 percentage points]; P < .001), but there was no significant difference in mean total cost of care ($3344 to $4087 vs $2977 to $3564; adjusted differential change, 1.0% [95% CI, -1.3% to 3.4%]; P = .39). In its first year, the 3PC population-based primary care payment system in Hawaii was associated with small improvements in quality and a reduction in PCP visits but no significant difference in the total cost of care. Additional research is needed to assess longer-term outcomes as the program is more fully implemented and to determine whether results are generalizable to other health care markets.

A home-based palliative care pilot program for patients with advanced cancer.

164 Background: Advanced cancer care often lacks a comprehensive approach; in Hawaii most oncology practices do not have access to palliative care teams. This causes high use of acute care services and suboptimal symptom management. Hawaii Medical Service Association (HMSA) created a pilot program called Supportive Care in which home-based palliative care services are offered to members with advanced cancer with goals of improved clinical outcomes and decrease utilization of acute care services Methods: Patients must have stage III or IV malignancy and ECOG PS of 2 or greater. Palliative care services are provided by Medicare-certified hospice agencies, with interdisciplinary teams, 24/7 on-call capacity, and expertise in symptom management. Hospice agencies are paid by HMSA to provide intermittent home visits. DME and pharmaceuticals. Care is coordinated with the patient's treating oncologist and other care providers. Each patient may receive 90 days of Supportive Care services in a 12 month period. Services are suspended during hospitalization or placement in a skilled nursing facility. Results: Patients enrolled in this program utilized hospital services significantly less than other Medicare Beneficiaries during the end-of-life period, suggesting that complex medical and psychosocial needs can be met in the home environment. The table compares findings from cancer patients enrolled in Supportive Care during 2014 with the most recent data available from The Dartmouth Atlas of Health Care. Conclusions: Multiple studies show improved quality of life for cancer patients receiving palliative care. Supportive Care resulted in improved clinical outcomes. Anecdotal feedback indicates high satisfaction among patients, families, and providers. Research to collect data and quantify satisfaction continues. [Table: see text]

The Economic Value of a Wellness and Disease Prevention Program

The objective of this study was to determine the economic impact of the Hawaii Medical Service Association's health promotion/disease prevention program. A retrospective analysis of health risk, health claims, and cost was performed using a mixed model factorial design for the years 2002-2005 that compared program participants to nonparticipants. All analyses were adjusted for preexisting observed differences based on sex, age, baseline morbidity, and health care costs between participants and nonparticipants using propensity score matching method and/or covariates as appropriate. After analyzing data from more than 166,000 HMSA members over a 4-year period, participants were found to incur consistently lower costs. Predictive modeling of upward cost trajectories relative to actual health care costs for participants and risk-matched nonparticipants indicated a savings of $350 per participant per year. Those who participated in additional wellness programming demonstrated additional cost savings. This study illustrates the economic value of a comprehensive health promotion program.

Monoamine oxidase inhibitor usage in Hawaii

Objective Previous studies involving prescriptions of monoamine oxidase inhibitors (MAOIs) have focused on predominantly Caucasian populations with little representation of Asian Americans and Pacific Islanders. The Asian American diet includes tyramine-rich fermented food items. This study describes the characteristics of MAOI prescribing patterns in Hawaii, a state with predominantly Asian Americans and Pacific Islanders. Methods Antidepressant usage including MAOIs were identified using a commercial insurance database from Hawaii Medical Service Association, a Blue Cross/Blue Shield subsidiary. Prescriptions from 1999 to 2003 were identified with basic information including ethnicity, age, diagnostic category, morbidity level, and 6 months adherence to MAOIs. Results Of the 28,890 patients prescribed antidepressants, seventeen individuals (0.06%, 95% confidence interval 0.03–0.09%) were prescribed MAOIs. MAOIs continue to be seldom used in treatment. There was no significant difference in prescriptions for MAOIs vs. other antidepressants based on ethnicity. Conclusions MAOIs are vastly underutilized in all ethnic groups including Asian American and Pacific Islander groups.

Pay-for-Performance Programs Show Positive Impact on Low-Performing Physicians

Pay-for-performance (P4P) programs are payment models that reward workers for meeting certain performance measures for quality and efficiency. In the health care setting, P4P programs use a variety of methods to reward physicians financially for achieving targets, including fee differentials and bonuses. A new study in the Journal for Healthcare Quality reveals that P4P programs seem to be effective in incentivizing low-performing physicians. These programs are increasingly being adopted by health plans and purchasers in the United States, the authors note. Over half of physicians practicing in commercial health maintenance organizations and about one quarter of all physicians in other settings participate in some form of P4P contracts, studies show. Still, as the new report points out, “only a few studies have assessed the ability of P4P programs to significantly improve alter physician behavior and the results are mixed.” Moreover, it remains unclear whether improvements would occur without the P4P program or whether they would be sustained. The new study supported by the Hawaii Medical Service Association in Honolulu investigated the impact of P4P in a Preferred Provider Organization (PPO) on low-performing physicians over a 4-year period. The study group consisted of providers in Hawaii who participated in a P4P program for ≥2 years from 1999 through 2007 and who qualified for awards on ≥1 of 7 clinical indicators: Angiotensin-converting enzyme inhibitor use among heart failure patients, mammography, cervical cancer screening, colorectal cancer screening, hemoglobin A1C testing for diabetes, the varicella vaccine, and the mumps, measles, and rubella vaccine. A comparison group of physicians consisted of providers enrolled in a large, commercial, PPO health plan for ≥2 years over the same time period as the Hawaii sample but not located in Hawaii and in a PPO without any performance-based incentive program. The groups were matched by specialty and patient quality indicator eligibility criteria. The P4P program financially incentivizes physicians in Hawaii (additional 1.5%–7.5% of their base professional fees) to perform quality care processes. Results showed that a P4P program in a PPO setting can improve quality of care. This was particularly the case for quality measures such as mammography, cervical cancer screening, A1C testing, and childhood immunization measures. Overall, the majority of quality scores were good, because more than three quarters of the eligible population received the recommended care in both groups. However, low-performing physicians using P4P improved significantly more than the comparison group. Further, although colorectal cancer screening had the lowest quality score, it showed the greatest improvement with time. In addition, the study found that the positive benefit of the P4P program may not be realized until the third or fourth year, highlighting the importance of sustaining P4P over longer periods of time. See: Journal for Healthcare Quality 2010;32:13–22.

Physical Activity Enjoyment Scale Short Form—Does It Fit for Children?

Click to increase image sizeClick to decrease image sizeKey words: invariancemultiethnic Additional informationNotes on contributorsRaheem J. PaxtonThis research was funded by the Hawaii Medical Service Association, an independent licensee of the Blue Cross and Blue Shield Association. Please address all correspondence concerning this article to Raheem J. Paxton, 140 Honness Lane Apt D, Ithaca, NY 14850.

Aligning Financial Incentives with Quality of Care in the Hospital Setting

This article describes the structure, implementation, and early results of a performance-based hospital incentive program designed by a large nonprofit health plan. The Hospital Quality Service and Recognition program, developed by the Hawaii Medical Service Association, was launched in 2001 to reward high-quality medical care at the hospital level. This pay-for-performance program used administrative claims data, survey data, and hospital-reported information to assess hospital performance in risk-adjusted complications and risk-adjusted length of stay (LOS), patient satisfaction, and hospital processes of care measures. Financial incentives were provided to participating hospitals based on their performance on these measures. Preliminary outcomes of the program evaluated over a 4-year period after implementation revealed improvements in aggregated rates of risk-adjusted surgical complications and efficiency of care as evidenced by a substantial decrease in risk-adjusted average LOS for several surgical procedures. Quality improvement was demonstrated in several other program components including emergency department satisfaction. This quality incentive program offers an innovative approach for encouraging delivery of high-quality and service-oriented care in a statewide network of participating hospitals.

A quality-driven physician compensation model: four-year follow-up study.

This case report describes a qualitative and preliminary quantitative assessment of a quality-based physician compensation program. The Hawaii Medical Service Association's Physician Quality and Service Recognition program offers an innovative and effective approach for improving delivery of high-quality and cost-effective care to patients enrolled in preferred provider organizations. Support for the program is demonstrated through increasing numbers of voluntarily participating physicians. Preliminary assessment of population outcomes reveals sustained improvements in many clinical areas and mixed findings in others. This study contributes to the body of knowledge available to payers and policy makers considering alternative payment methods to reward improved performance.

Antibacterials for Viral Respiratory Infections

To determine the effectiveness of a multi-component managed-care intervention in reducing inappropriate antibacterial prescribing for viral respiratory infections (VRIs). An episode-based retrospective analysis of administrative claims data between 1998, the year prior to the initiation of the first component of the intervention, and 2001, the year after full implementation. Patient group studied: All health plan members of the Hawaii Medical Service Association (Blue Cross Blue Shield Hawaii) in the US who sought care for an upper respiratory infection. To ensure that changes in the rate of antibacterial prescriptions were not because of viral-related infections being coded as bacterial, we measured, in parallel, diagnoses of similar bacterial respiratory infections. Persons with chronic bronchitis, emphysema, and asthma were excluded. Episodes temporally associated with any diagnosis for a non-respiratory bacterial infection were also excluded. The main study outcome was change in VRI-associated antibacterial utilization. Secondary outcomes included episode-related outpatient, emergency department and inpatient utilization, and estimated cost savings related to changes in antibacterial prescribing rates. The population-based rate of antibacterial use for VRIs decreased significantly between 1998 and 2001, from 39.9% to 33.5%. There was no observed increase in the proportion of bacterial to viral episodes over the same period of time, suggesting that the observed decrease in rates was not a result of coding changes. This reduction in antibacterial prescribing for VRIs was not associated with an increase in emergency department or inpatient utilization. A reduction in VRI-related outpatient visits was observed. The reduction in inappropriate antibacterial prescriptions was estimated to result in a cost saving of approximately $US500 000 over the 4-year study period, assuming the antibacterials prescribed were similar over time (1998 values). The majority of savings were found in the last year, after full implementation of the intervention. While this study did not offer evidence that one component of the intervention was effective by itself, the combination of all components was associated with decreased antibacterial prescribing for VRIs. The observed reduction in antibacterial prescribing for viral episodes was not associated with increased healthcare visit utilization, and is estimated to be associated with a sizable cost saving. While further research is recommended, this study offers preliminary evidence that multi-component physician feedback programs may be effective in reducing the inappropriate prescribing of antibacterials, and managed care organizations can use claims-based, episode-defined analyses to evaluate the effectiveness of their initiatives.

Medical costs of managed care in patients with type 2 diabetes mellitus

The main objective of this study was to analyze the principal treatment cost drivers in patients with type 2 diabetes mellitus in a managed care setting. The study used retrospective integrated (linked) medical and pharmacy claims data for the calendar year 1995. The data were obtained from, and in cooperation with, the Hawaii Medical Service Association, Honolulu, Hawaii. The medical claims data included paid claims for services and procedures for diabetes and commonly associated comorbidities. Claims and associated costs for pharmacotherapy administered to the patient population were recorded in the pharmacy data. Patients aged ≥65 years were excluded because Medicare claims were unavailable for the type 2 diabetic population. The sample used in this study included 5171 patients. An ordinary least squares regression model was employed to identify principal cost drivers among the identified cohort to the managed care system. Independent variables in the analysis consisted of the presence or absence of a number of commonly observed comorbidities associated with diabetes mellitus (hypertension, hyperlipidemia, cardiovascular diseases, congestive heart failure, renal disorders, retinopathy, neurologic disorders, and any cardiac or noncardiac comorbidity combinations), pharmacologic therapy variables (insulin, oral medication, or both), a number of significant events (hospitalization, dialysis, hemoglobin A 1c testing, and eye examination), patient enrollment category (fee-forservice vs a capitated system), and patient age and sex. The dependent variable was the natural logarithm of total medical costs of treatment for diabetes and commonly observed comorbidities. Results showed that among comorbidity variables, the 3 largest treatment cost drivers for patients with type 2 diabetes were the presence of neurologic disorders, renal disorders, and any comorbidity combination (cardiac or noncardiac or both), in decreasing order of significance. Similarly, higher costs of treatment were associated with episodes of hospitalization, use of antidiabetic medication, dialysis services, and hemoglobin A 1c testing. Whether the patient was being treated under a capitated provider payment system or a fee-for-service system did not have any significant impact on the medical costs of diabetes-related treatment. Age was positively associated with these costs, indicating that older patients were more likely to incur higher costs to the system. The overall explanatory power of the model was 40%. In summary, unless diabetes is properly managed and glucose levels monitored, some component of an integrated health system (hospital vs pharmacy) necessarily bears financial risk. An understanding of the underlying cost distribution for a chronic disease could help in targeting interventions, integrating disease-management services, and managing the formal structure of the health plan being considered.

PGD5 Predicting Hospitalization of Patients with Diabetes Mellitus: An Application of the Bayesian Discriminant Analysis

OBJECTIVE: The main objectives of this paper were to develop and subsequently test a Bayesian discrimination model for the purpose of identifying both the personal and the health care system characteristics predictive of hospitalization of the treatment of diabetes or commonly observed comorbidities associated with the disease. METHODS: First, a Bayesian classification framework was proposed to discriminate patients into two groups-the high risk and the low risk. The model was then tested by using a logit regression technique in order to estimate the probability of one or more hospitalization events among diabetes patients. The study used claims data extracted from the Hawaii Medical Service Association (HMSA) Private Business Claims (PBS) files for the calendar year 1995. Patients under 65 years were identified by paid claims with ICD-9-CM diagnosis codes of 2.50.xx which gave a sample size of 6,841. RESULTS: Age, gender, various pharmacotherapy variables, presence of hypertension, hyperlipidemia, coronary heart failure, multiple cardiovascular diseases, any combination of commonly observed comorbidities, dialyis services, and annual eye examination were highly predictive of one or more hospitalization events. The model showed a predictive power of almost 90% CONCLUSION: Multivariate discriminant analysis using a logit regression model successfully (1)identified important explanatory variables predictive of hospitalization, (2) assigned patients into one of two mutually exclusive classes, and (3) offered a benchmark for a comprehensive disease management strategy for more involved diabetic patients.

Predicting hospitalisation of patients with diabetes mellitus. An application of the Bayesian discriminant analysis.

The objective of this study was to develop, and subsequently test, a Bayesian discrimination model for the purpose of identifying both the personal and the healthcare system characteristics predictive of hospitalisation for the treatment of patients with diabetes mellitus or commonly observed cormorbidities associated with the disease. First, a Bayesian classification framework was proposed. The model was then tested by using a logit regression technique in order to estimate the probability of one or more hospitalisation events among patients with diabetes. The study used claims data extracted from the Hawaii Medical Service Association (HMSA) Private Business Claims (PBS) files for the 1995 calendar year. Patients under 65 years were identified by paid claims with ICD-9-CM diagnosis codes of 250.xx which gave a sample size of 6841 patients. Age, gender, various pharmacotherapy variables, presence of hypertension, hyperlipidaemia, congestive heart failure, multiple cardiovascular diseases, any combination of commonly observed comorbidities, dialysis services and annual eye examination are highly predictive of 1 or more hospitalisation events. The model shows a predictive power of almost 90%. This study found that multivariate discriminant analysis using a logit regression model successfully identifies: (i) important explanatory variables predictive of hospitalisation; (ii) assigns patients into 1 of 2 mutually exclusive classes; and (iii) offers a benchmark for a comprehensive disease management strategy for patients with more complicated diabetes.

Use of Glycosylated Hemoglobin Testing Among Patients with Diabetes Mellitus

OBJECTIVE: To estimate the likelihood of the use of glycosylated hemoglobin testing (HbA1c) among a cohort of patients with diabetes mellitus in a managed care setting. DESIGN: A retrospective, cross-sectional analysis of managed care claims data obtained from the private business claims (PSC) files of the Hawaii Medical Service Association for the calendar year 1995. A probit model was used to estimate the probability of the dependent variable (HbA1c). Independent variables specified in the model included age, gender, presence of common comorbidities, use of insulin and oral anti-diabetic agents, and type of insurance plan. The probit model was run for an n of 6,841 patients. SETTING: Health insurer in the state of Hawaii. Includes managed care and indemnity products. PATIENTS: 6,841 patients under 65 years old were identified by paid claims with ICD-9-CM codes of 250.xx. PARTICIPANTS: Not applicable. INTERVENTION: Not applicable. MAIN OUTCOME MEASURES: Use of HbA1c testing. RESULTS: Age and gender are not statistically significant as predictors of HbA1c testing. Patients with multiple comorbid conditions, hyperlipidemia, and multiple cardiovascular comorbidities have a higher probability of receiving HbA1c testing relative to the diabetic population with no comorbidities. Patients receiving services under an HMO plan and patients on polypharmacy (oral agents plus insulin) are more likely to receive an HbA1c test. The model demonstrated an explanatory power of 64.05%. CONCLUSION: Patients with greater clinical involvement were found to be more likely to receive HbA1c testing over patients with less disease involvement. Patients enrolled in an HMO plan are more likely to receive HbA1c testing relative to patients enrolled in fee for-service (FFS) plans. Finally, the probit model demonstrated a high level of explanatory power. This is an important criterion if we are to place some level of confidence in claims data for: a) obtaining baseline information; b) planning purposes; and c) periodic evaluation of diabetes care interventions.