The Impact of Reimbursement for Non–Face-to-Face Chronic Care Management on Comprehensive Metabolic Biomarkers Among Multimorbid Patients With Type 2 Diabetes

Diabetes mellitus, especially type 2 diabetes mellitus, is a leading public health concern in the United States with an estimated total cost of $327 billion in 2018.1 The prevalence of diabetes in the United States is one of the highest prevalence in the world.2 Approximately 10% of the US population have type 2 diabetes.3 Patients and clinicians frequently face various challenges to managing type 2 diabetes, such as poor glycemic control, the presence of comorbidities, and complications.4,5 In addition, patients usually face barriers to seeking chronic care, such as lack of appointment times, long waiting times, or distance from provider locations.6 To better manage diabetes, various approaches have been developed and implemented to improve care.7,8

Chronic care management (CCM) for aging populations represents an important approach to addressing disease management for the US population, where care is increasingly provided out of clinical settings through telephonic or video visits. In 2015, the Centers for Medicare and Medicaid Services (CMS) introduced a new CCM payment policy and began reimbursing non–face-to-face CCM (NFFCCM) services for Medicare beneficiaries who have multiple (2 or more) chronic conditions, including diabetes, which aims to provide better primary care coordination and disease management.9 By the end of 2017, 3 billing codes were issued, Current Procedural Terminology (CPT) codes 99490, 99487, and 99489. The code of 99490 was issued in 2015 for noncomplex CCM and the other 2 supplementary billing codes were released in 2017 for complex CCM.9

Compared with in-person care, earlier evidence indicated that NFFCCM is a successful alternative to reduce health care costs among Medicare beneficiaries, likely through decreased use of inpatient hospital and postacute care services.10 Other related chronic care models have already demonstrated success in helping diabetes patients maintain or improve their health by improved glycemic control compared with traditional chronic care models.8,11,12 With a high proportion of Medicare beneficiaries with diabetes, the initiation of reimbursement for NFFCCM represents an opportunity to improve diabetes care management and outcomes. However, no studies have been published to examine the impacts of NFFCCM on diabetes management, especially on glycemic control and other clinical risk factors, such as body mass index (BMI), blood pressure (BP), and low-density lipoprotein cholesterol (LDL-C). Therefore, we conducted this quasiexperimental study to examine the impacts of NFFCCM services, complex or noncomplex, on glycemic control and other clinical risk factors results among Medicare beneficiaries with type 2 diabetes in Louisiana.

MATERIALS AND METHODS Study Design and Data Sources

This is a quasiexperimental study using electronic health records (EHRs) between January 2013 and February 2020 from Research Action for Health Network (REACHnet), a clinical research network in PCORnet, the National Patient-Centered Clinical Research Network. Records stored in REACHnet are from several health systems in Louisiana and Texas and standardized to the PCORnet Common Data Model. The study and analysis plan were approved by Tulane University Institutional Review Board (IRB# 906810).

Inclusion and Exclusion Criteria

All type 2 diabetes patients from 3 health systems in Louisiana were identified using the Surveillance Prevention, and Management of Diabetes Mellitus (SUPREME-DM) definition,13 which provides detailed guidelines to select diabetes patients when using EHR data. We then included patients with Medicare as their primary payer. We excluded type 1 diabetes and gestational diabetes in this study. We excluded patients with missing baseline characteristics or no visits in the postperiod for statistical analysis. The baseline and postperiods were defined in one of our published studies.14

Outcome Measures

The primary outcome measure is glycemic control measured by glycated hemoglobin (HbA1c). HbA1c measurements included: mean values of HbA1c during follow-up; whether mean values HbA1c were <9%, 8%, or 7%; whether the decline of HbA1c levels from baseline was ≥0.5%, 0.3%, 0.2%, 0.1%. The secondary outcomes included BMI, BP, and LDL-C levels in the follow-up period. Outcomes related to BMI included mean values of BMI and proportions of each BMI category (underweight, normal weight, overweight, obese).15 The outcomes related to BP included mean values of diastolic BP and systolic BP and proportions of each BP category (normal, elevated, stage 1 hypertension, and stage 2 hypertension).16 LDL-C was measured as mean values of LDL-C during follow-up; whether mean values of LDL-C was <100, 110, or 130 mg/dL; and whether the LDL-C decline from baseline was ≥15, 10, or 5 mg/dL.

Statistical Analysis

We defined the treatment group as patients with at least 1 record of NFFCCM (CPT 99490, 99487, and 99489), and the dates of the first NFFCCM coded were the initiation dates. We then assigned initiation dates randomly for untreated patients based on the distribution of initiation dates in the treated population. The baseline period was 24 months before the initiation dates. we assessed the outcomes in 24 months after the initiation dates. To find a group of beneficiaries resembling NFFCCM beneficiaries, we firstly used the group-based trajectory modeling to divide our sample into several groups who shared a similar trend of outpatient visits at baseline by controlling a set of observable characteristics including age, race, ethnicity, chronic conditions, utilization, and several biomarkers related to diabetes. The selection of the trajectory model was documented and applied in previous studies.17–20 We then additionally included indicators of trajectory groups with baseline characteristics in the propensity score weighting process to obtain a successful balance between treatment and control groups. The propensity score was estimated from a probit regression model fitted on our analytic sample that includes both NFFCCM beneficiaries and non-NFFCCM beneficiaries. NFFCCM beneficiaries were assigned a weight of 1 and non-NFFCCM beneficiaries were assigned weights based on propensity scores [weight=pscore/(1-pscore)]. These weights were used in our subsequent outcome modeling. We used this sample in a weighted linear regression to implement the doubly robust estimator by controlling the same set of variables used in the weighting step.

In addition to assessing the marginal association between ever having received NFFCCM and study outcomes during the 24-month follow-up, we also conducted multiple subgroup analyses based on the complexity of NFFCCM (CPT 99487 and 99489) and different baseline levels of HbA1c, BMI, BP, and LDL. Lastly, we tested different treatment definitions based on different treatment periods (6 months or 12 months) from the initiation and frequency of NFFCCM use in the treatment periods. The outcomes were collected in the remaining 18 months for 6-month treatment period and remaining 12 months for 12-month period. We then divided the treated patients into groups based on the frequency of NFFCCM use during the treatment period. The treatment sample may change with different treatment definitions; therefore, we repeated our weighting process and regenerated propensity scores for non-NFFCCM beneficiaries to approximate the corresponding counterfactuals. For every weighting, we checked the standardized mean difference between treatment and control before and after weighting to ensure successful balance defined as differences within 10% for all baseline characteristics. All analyses were performed using SAS 9.4 and Stata 15.1.

RESULTS

We identified 19,025 type 2 diabetes mellitus patients with baseline and follow-up HbA1c tests. Totally, 1501 were NFFCCM beneficiaries and 17,524 were non-NFFCCM beneficiaries. The baseline characteristics of this sample are shown in Table 1. We obtained 3 groups by using group-based trajectory modeling and patients in each group shared a similar trend of outpatient visits during 24 months at baseline (Supplemental Fig. A1, Supplemental Digital Content 1, https://links.lww.com/MLR/C586). All baseline characteristics were successfully balanced within 10% of a standardized mean difference after weighted by propensity scores. In the matched sample, the mean age was about 72 years old and nearly half were Black. Ninety-six percent had any diagnosis record of hypertension. Fifteen percent had baseline HbA1c over 8%. We then repeated weighting procedures and successfully matched all characteristics for the treatment and control groups in each sample listed in the following outcome tables.

TABLE 1 - Baseline Characteristics Before and After Propensity Score Weighting for HbA1c Measurements Nonweighted (%) Weighted (%) Baseline characteristics Treatment Control SMD Treatment Control SMD Age at first CCM (y) 72.42 72.434 0.1 72.42 72.436 0.2 Female 59.4 54.3 −10.3 59.4 59.2 −0.4 Black 47.3 38.4 −18.0 47.3 47.2 −0.1 Hispanic 1.7 2.6 6.3 1.7 1.7 −0.1 Stroke 15.3 14.9 −1.2 15.3 15.2 −0.3 Hypertension 96.0 95.6 −2.1 96.0 96.1 0.6 Alzheimer’s 1.7 1.4 −2.6 1.7 1.7 −0.2 Arthritis 51.8 46.6 −10.5 51.8 51.8 0.0 Asthma 14.5 11.2 −10.1 14.5 14.5 −0.1 Atrial fibrillation 14.3 13.8 −1.4 14.3 14.3 0.2 Cancer 12.4 12.3 −0.3 12.4 12.4 0.0 Chronic obstructive pulmonary disease 25.1 22.2 −6.9 25.1 25.0 −0.2 Chronic kidney disease 45.0 45.0 0.1 45.0 44.9 −0.1 Depression 26.4 23.1 −7.6 26.4 26.1 −0.5 Heart failure 22.3 20.2 −5.1 22.3 22.1 −0.4 Hyperlipidemia 90.3 89.6 −2.6 90.3 90.3 0.0 Coronary heart disease 35.6 34.5 −2.4 35.6 35.4 −0.5 Osteoporosis 13.5 11.6 −5.9 13.5 13.5 0.0 Outpatient visits per month 0.619 0.549 −13.2 0.619 0.618 −0.2 Emergency department visits per month 0.018 0.016 −4.9 0.018 0.017 −0.2 Have any inpatient visits 24.2 22.3 −4.4 24.2 24.1 −0.2 Number of A1c tests 4.252 3.861 −21.3 4.252 4.259 0.4 Mean A1C >8% 15.5 15.9 1.3 15.5 15.3 −0.3 BMI ≥35 (kg/m2) 31.2 27.5 −8.2 31.2 31.4 0.5 LDL ≥110 (mg/dL) 19.3 22.7 8.2 19.3 19.3 −0.1 HDL <40 (mg/dL) 53.9 53.6 −0.6 53.9 53.8 −0.2 eGFR ≥30 (mL/min/1.73 m2) 95.9 95.0 −4.3 95.9 95.9 −0.1 Trajectory group 1 30.0 36.1 12.9 30.0 30.0 0.0 Trajectory group 2 55.8 53.9 −3.9 55.8 55.9 0.2 Trajectory group 3 14.2 10.0 −12.7 14.2 14.1 −0.3 N 1501 17,524 1501 17,524

BMI indicates body mass index; CCM, chronic care management; eGRF, estimated glomerular filtration rate; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; LDL, low-density lipoprotein; SMD, standardized mean difference.


Glycated Hemoglobin

We first compared patients using any NFFCCM services with patients without NFFCCM on several HbA1c measurements in 24 months from baseline (Table 2). NFFCCM beneficiaries experienced better HbA1c management. Receiving any NFFCCM reduced the mean values of HbA1c by 0.063% (95% CI: 0.031%–0.094%; P<0.001). The proportion of NFFCCM beneficiaries with HbA1c <7% was 1.6% (95% CI: 0.3%–2.9%; P=0.013) higher than that among non-NFFCCM beneficiaries. We additionally evaluated the association between NFFCCM and HbA1c by different complexity (Table 2). Receiving any complex NFFCCM was associated with an increase in the proportion of HbA1c <8% of 3.1% (95% CI: 2.1%–4.0%; P<0.001). The magnitude of the benefit of receiving only noncomplex NFFCCM on each HbA1c measurement was similar to that of receiving any NFFCCM.

TABLE 2 - The Effect of NFFCCM on HbA1c Outcomes Receiving any CCM Receiving any complex CCM Receiving only noncomplex CCM Baseline HbA1c ≥7.5% Baseline HbA1c <7.5% Mean HbA1c (%)* −0.063*** 0.006 −0.069*** −0.151*** −0.017 [−0.094,−0.031] [−0.026,0.038] [−0.101,−0.038] [−0.233,−0.069] [−0.046,0.012] 0.000 0.723 0.000 0.000 0.240 Proportion of HbA1c <9%† 0.015*** −0.009* 0.018*** 0.060*** 0.000 [0.009,0.022] [−0.017,−0.002] [0.011,0.024] [0.037,0.084] [−0.004,0.004] 0.000 0.017 0.000 0.000 0.989 Proportion of HbA1c <8%‡ 0.022*** 0.031*** 0.021*** 0.059*** 0.005 [0.013,0.031] [0.021,0.040] [0.012,0.030] [0.031,0.087] [−0.002,0.013] 0.000 0.000 0.000 0.000 0.157 Proportion of HbA1c <7%§ 0.016* −0.005 0.019** 0.016 0.007 [0.003,0.029] [−0.019,0.008] [0.006,0.031] [−0.006,0.037] [−0.008,0.021] 0.013 0.442 0.004 0.153 0.363 Proportion of HbA1c decline ≥0.5%‖ −0.011* 0.004 −0.014* 0.044** −0.026*** [−0.022,−0.001] [−0.008,0.017] [−0.024,−0.003] [0.015,0.072] [−0.036,−0.015] 0.039 0.488 0.013 0.003 0.000 Proportion of HbA1c decline ≥0.3%¶ −0.009 −0.005 −0.010 0.040** −0.020** [−0.021,0.004] [−0.019,0.010] [−0.023,0.002] [0.011,0.069] [−0.033,−0.006] 0.181 0.533 0.114 0.007 0.005 Proportion of HbA1c decline ≥0.2%# 0.003 0.007 0.001 0.058*** −0.010 [−0.011,0.016] [−0.008,0.023] [−0.013,0.015] [0.029,0.087] [−0.026,0.005] 0.691 0.336 0.884 0.000 0.186 Proportion of HbA1c decline ≥0.1%†† −0.007 −0.001 −0.009 0.048** −0.021* [−0.021,0.007] [−0.017,0.014] [−0.023,0.006] [0.019,0.076] [−0.037,−0.005] 0.320 0.868 0.239 0.001 0.011 N treatment 1501 156 1345 358 1143 N control 17,524 16,333 17,549 4431 13,052 N total 19,025 16,489 18,894 4789 14,195

*Mean HbA1c during the 2-year follow-up period.

†The percent of patients with mean HbA1c <9%.

‡The percent of patients with mean HbA1c <8%.

§The percent of patients with mean HbA1c <7%.

‖The percent of patients whose HbA1c was declined by 0.5% or more than 0.5% from baseline.

¶The percent of patients whose HbA1c was declined by 0.3% or more than 0.3% from baseline.

#The percent of patients whose HbA1c was declined by 0.2% or more than 0.2% from baseline.

††The percent of patients whose HbA1c was declined by 0.51% or more than 0.1% from baseline. For each outcome, the coefficient was listed and followed with its 95% CI and P value.

CCM indicates chronic care management; HbA1c, glycated hemoglobin; NFFCCM, non–face-to-face chronic care management.

*P<0.05.

**P<0.01.

***P<0.001.

We then explored the associations between receiving any NFFCCM and HbA1c values by different levels of baseline HbA1c (columns 4 and 5 in Table 2). For those who had higher HbA1c (≥7.5%) at baseline, receiving any NFFCCM was associated with a reduction in HbA1c of 0.151% (95% CI: 0.069%–0.233%; P<0.001) and an increase in the proportion with HbA1c <9% or 8% of 6.0% and 5.9%, respectively (P<0.001). A higher proportion of patients had an HbA1c reduction from baseline among NFFCCM beneficiaries. For example, the proportion of NFFCCM beneficiaries with an HbA1c reduction of at least 0.2% is about 5.8% (95% CI: 2.9%–8.7%; P<0.001) higher than that among non-NFFCCM beneficiaries. For those who had lower HbA1c (<7.5%) at baseline, however, we found no significant improvement in glycemic control related to receiving any NFFCCM services.

We repeated our analyses to test whether the frequency of noncomplex NFFCCM would change those associations (Supplemental Table A1, Supplemental Digital Content 1, https://links.lww.com/MLR/C586). Two treatment periods were defined, 6-month and 12-month from the first NFFCCM. Outcomes were collected in the following months after the treatment period. During 6-month treatment, about 47% of noncomplex NFFCCM beneficiaries received the services more than 5 times. We only found a higher frequency of noncomplex NFFCCM was associated with a significant improvement in the proportion of patients with an HbA1c decline of at least 0.1%, an increase of 6.1% (95% CI: 0.4%–11.9%; P=0.036). About 46% of patients received noncomplex NFFCCM more than 10 times in 12 months. Receiving noncomplex NFFCCM over 10 times was associated with a reduction in HbA1c of 0.143% (P=0.086).

Body Mass Index

We examined the impact of receiving NFFCCM services on BMI changes (Table 3). During 24-month follow-up from the first NFFCCM service, receiving any NFFCCM was associated with a reduction in the mean BMI values of 0.155 kg/m2 (95% CI: 0.029–0.282 kg/m2; P=0.016), an increase in the proportion of patients with a normal BMI range of 1.2% (95% CI: 0.2%–2.1%; P=0.015), and a decrease in the overweight category of 1.6% (95% CI: 0.5%–2.8%; P=0.006). For patients with baseline BMI <35 kg/m2, receiving any NFFCCM services was associated with a reduction in mean BMI of 0.205 kg/m2 (95% CI: 0.076–0.334 kg/m2; P=0.002), an increase in normal category of 1.8% (95% CI: 0.5%–3.1%; P=0.005), and a decrease in the overweight category of 2.2% (95% CI: 0.6%–3.8%; P=0.006). We found no significant dose-response effects of receiving NFFCCM on BMI changes (Supplemental Table A2, Supplemental Digital Content 1, https://links.lww.com/MLR/C586).

TABLE 3 - The Effect of NFFCCM on BMI Outcomes Receiving any CCM Receiving any Complex CCM Receiving only Noncomplex CCM Baseline BMI ≥35 kg/m2 Baseline BMI <35 kg/m2 BMI (kg/m2) −0.155* 0.074 −0.192** −0.081 −0.205** [−0.282,−0.029] [−0.068,0.216] [−0.318,−0.066] [−0.375,0.214] [−0.334,−0.076] 0.016 0.308 0.003 0.591 0.002 Proportion of underweight 0.001 0.001 0.001 0 0.002 [−0.001,0.004] [−0.001,0.004] [−0.001,0.004] [−0.001,0.000] [−0.001,0.005] 0.316 0.343 0.263 0.373 0.231 Proportion of normal 0.012* 0.006 0.013** 0 0.018** [0.002,0.021] [−0.004,0.015] [0.003,0.022] [−0.002,0.001] [0.005,0.031] 0.015 0.246 0.008 0.37 0.005 Proportion of overweight −0.016** −0.002 −0.017** −0.003 −0.022** [−0.028,−0.005] [−0.014,0.010] [−0.029,−0.006] [−0.007,0.001] [−0.038,−0.006] 0.006 0.722 0.003 0.094 0.006 Proportion of obese 0.003 −0.005 0.004 0.004 0.002 [−0.008,0.014] [−0.016,0.006] [−0.007,0.015] [−0.000,0.008] [−0.013,0.017] 0.542 0.408

留言 (0)

沒有登入
gif