Abstract
Objective
To evaluate the effect of a forced disruption to Medicaid managed care plans and provider networks on health utilization and outcomes for children with persistent asthma.
Data Sources
Medicaid managed care administrative claims data from 2013 to 2016, obtained from a southeastern state.
Study Design
A difference‐in‐difference analysis compared patients’ outpatient, inpatient, and emergency department (ED) utilization and receipt of recommended services before and after implementation of a statewide redistribution of patients among nine managed care plans.
Data Collection/Extraction Methods
Enrollment data for children with asthma were linked to the administrative claims. Children were included if they had a diagnosis of persistent asthma in 2013 and if they were enrolled continuously throughout 2014‐2016.
Principal Findings
Among the 28 537 children with asthma, 26% were forced to switch their managed care plan after the redistribution. Of these, 67% also switched their primary care provider (PCP). Relative to those who remained in their plan, disruption was associated with an additional 2.1 percentage‐point decrease in the number of children who had an outpatient visit per quarter [95%CI −2.8, −1.3], from 71% to 66% (compared to plan stayers: 74% to 71%). Among children experiencing a change to their plan, there was overall a decrease in the proportion of children receiving an asthma‐specific visit per quarter, but there was less of a decrease in children that also changed their PCP [1.6 percentage points, 95%CI 0.7, 2.5], from 9.7% to 8.3% (compared to those who did not switch their PCP: 12% to 8.6%). Indicators of asthma care quality and emergent care utilization were not significantly different between the two periods.
Conclusions
While there was a decrease in the number of outpatient visits associated with forced disruption of Medicaid managed care plans for children with persistent asthma, there were no consistent associations with worse asthma quality performance or higher emergent health care utilization.
Keywords: asthma, child, disruptions to care continuity, managed care programs, Medicaid
What This Study Adds.
What is already known on this topic?
Prior observational studies suggest disruptions to care are associated with worse health outcomes.
Observational studies struggle to differentiate between causality and association in the relationship between disruptions and health.
No assessment has been conducted leveraging a natural experiment to examine the effects of managed care plan disruptions on quality and utilization in children.
What do we know now?
For children with asthma continuously enrolled in Medicaid, there was a greater decrease in outpatient visits for those forced to switch their managed care plan, relative to those who did not switch.
There do not appear to be negative effects on quality and emergent utilization for this population attributable to the forced switch in managed care plan or primary care provider.
Continuous access to care may be more important to the intermediate‐term health of children with asthma enrolled in Medicaid than continuity of care with a single plan or provider.
1. INTRODUCTION
Managed care is the dominant model of financing and delivery for over 70 million Americans who receive health insurance coverage through Medicaid. 1 In 2016, approximately 70% of all children covered by Medicaid were enrolled in managed care plans that use networks of providers to coordinate comprehensive care needs for those with complex conditions. 2 However, contracts between state Medicaid agencies and managed care plans end frequently, often requiring enrollees to be transitioned to new plans. 3 While Medicaid plans operating in the same state offer standard benefits and formularies, they vary widely in their approach to facilitating care—including quality improvement strategies, aggressiveness of utilization review, and physician network stability—all of which influence how recipients access care. 4 , 5 , 6 Moreover, there is typically limited overlap between the provider networks of managed care plans. 7 Consequently, forced changes in plans often trigger changes in primary care providers (PCP) for recipients, leading to potential disruptions in continuity of care.
The potential effects of discontinuities in care are particularly relevant for low‐income children with chronic conditions, who are often exposed to more intrinsic and extrinsic risk factors for poor disease control than higher‐income children. 8 Similarly, understanding the effect of discontinuities is of great importance for state Medicaid programs, who must determine whether to renew contracts with managed care plans. Yet, findings from research broadly examining the effects of such plan discontinuities have previously been mixed, with some studies showing increased emergency department (ED) or inpatient use, and others showing decreased access to health care or no effect on patient utilization. 9 , 10 , 11 , 12 Establishing a causal link between discontinuity and patient outcomes has been difficult because changes in plans or providers may be initiated by patients, introducing a potential for selection bias.
In January 2015, a southeastern state reorganized their Medicaid managed care configuration from six to nine regional plans to more evenly distribute enrollees across plans. During this implementation, almost one‐third of the state's managed care enrollees were selected to be transferred to different health plans. We exploit this natural experiment to estimate the causal impact of involuntary disruptions of plan, and subsequent provider network changes, on children with persistent asthma.
2. METHODS
2.1. Sources of data and study population
We obtained Medicaid administrative claims data from the southeastern state for all managed care recipients from 2013 to 2016. The data included person‐level administrative enrollment files linked to health insurance claims from the regional managed care plans participating in the state's Medicaid program. This study was approved by the institutional review board at Yale University.
Our study population consisted of all children between the ages of 5 to 18 during our study period who were continuously enrolled in the state's Medicaid program (ie, no more than a 45‐day gap in enrollment per year) from 2014 to 2016, and who had a diagnosis of persistent asthma in 2013. Consistent with standard modifications to the National Committee for Quality Assurance (NCQA)’s Health Effectiveness Data and Information Set (HEDIS), we defined patients with persistent asthma as those who met one of the following criteria: (a) at least one ED visit or hospitalization with a primary diagnosis of asthma, (b) at least three outpatient visits with a primary diagnosis of asthma, or (c) a prescription for at least one asthma controller medication. 13 Children under the age of five were not included in the study due to inconsistency in asthma diagnosis at this younger age. 14 , 15
We identified each enrollee's assigned PCP by a variable supplied in the state's Medicaid enrollment records. PCPs could be either MD/DOs, nurse practitioners, or physician assistants. To obtain the final analytic sample, we identified a group of “plan switchers,” defined as patients in our study period who switched plans in January 2015; our corresponding control group of “plan stayers” were patients who were continuously enrolled in a single health plan for 2014‐2016. Among the plan switchers, we identified a group of “PCP switchers,” defined as patients who also switched PCPs in January 2015, and at no other time during our study period. We limited our switcher sample to patients who switched plans and/or PCPs in the month of January, the month of the initial transition (Appendix S1: Figure S1 and Table S5). The corresponding control group within plan switchers, labeled “PCP stayers,” were patients who remained with their PCP during our study period. In this January 2015 reshuffling, all patients who switched plans transitioned from an established plan in the area to a new regional plan entrant (Appendix S1: Table S6).
2.2. Variables
Our primary outcomes of interest were health care utilization and quality measures specific to children with persistent asthma. We assessed health care utilization based on the proportion of enrollees with an outpatient, ED, or inpatient care visit in a given quarter. We also included measures of these services that were specifically related to asthma, where the primary diagnosis on the claim was asthma.
We measured quality of care using two HEDIS pediatric asthma‐related quality measures. The first is a measure of asthma medication adherence, or the percentage of children with persistent asthma who remained on an asthma controller medication for at least 50% of a given year. The second is the asthma medication ratio, defined as the percentage of children with persistent asthma whose percentage of total asthma medication prescriptions dispensed as controller medications was at least 50% in a given year. 15 For each measure, we used the technical specification published by The Centers for Medicare and Medicaid Services (CMS) to construct the measure from enrollee claims data at the enrollee‐year level (2014‐2016). 15
To assess the differences in medical complexity between plan and PCP switchers/stayers, we used the Van Walraven comorbidity score, a standard modification derived from the Elixhauser comorbidity index that has been used in research involving administrative data for pediatric populations and accounts for patient comorbidities. 16 , 17 , 18 , 19 , 20 , 21 , 22
2.3. Statistical analysis
We used a two‐step difference‐in‐difference approach to first estimate the causal effect of a forced plan switch among children with persistent asthma and to then estimate the causal effect for plan switchers that also switched PCPs, comparing outcomes in the preperiod (2014) to the postperiod (2015‐2016), and between “stayers” and “switchers.” We chose to keep our preperiod to one year to reduce attrition given the need for continuous enrollment in the sample requirement and our postperiod to two years to capture longer term potential changes in utilization and quality.
A key methodological assumption in our causal framework is that switches resulting from the January 2015 legislation are involuntary, such that plan/PCP switching is not the result of an unobserved mechanism that might also be correlated with the outcomes and thus bias our estimated effects. Within patient groups, we tested for parallel trends in the preswitch period between the plan and PCP switcher groups for all outcomes and found no evidence of any significant change in preswitch trends between cohorts. Detailed results for the parallel trends test are given in the Appendix S1: Tables S1‐S4. We used the Holm‐Bonferroni correction for multiple comparisons to reduce the likelihood of type I errors. 23 We provide additional methodological details of the statistical approach and the Holm‐Bonferroni correction in the Appendix S1.
We included patient‐level fixed effects to control for both observable and unobservable differences in patient characteristics. We added quarter and year fixed effects to address the seasonality of asthma exacerbations in the spring and early fall. 24 , 25 , 26 , 27 County‐level fixed effects for the county of a patient's residence were included to control for changes in local area resources, responding to the fact that a small portion of our patient population moves during the study period, and account for the geographic variations in asthma outcomes and utilization patterns. 28 , 29 , 30 , 31 Finally, we clustered standard errors at the patient level.
We conducted five additional analyses to test for the robustness of our results. The first was a subgroup analysis that focused on a subset of patients that we believed would be particularly affected by the forced discontinuity, namely, patients who were “connected to care” in 2014. These patients were defined as those who had at least one visit with their assigned PCP in the calendar year prior to the transition. The second was a bounding exercise that sought to address any concern that our analysis included patients who were voluntary switchers (thus, a violation to our causal inference assumption that switches are involuntary). In this analysis, we dropped patients with the most extreme changes in outcomes based on baseline plan/PCP switch rates to determine if we would still obtain the same results absent these “extreme” effects. In the third and fourth analyses, we used patient controls instead of patient fixed effects and Area Health Resources File (AHRF) county‐level measures (num. of MDs per 100 000 residents [2015] and num. of hospital beds per 100 000 residents [2014]) instead of county‐fixed effects to determine if we were over‐controlling for, and thus missing, any relevant variation. In the fifth analysis, we used the number of visits (outpatient, ED, inpatient) as our primary outcome as opposed to ever any visit. Additional details of these methods and results can be found in the Appendix S1: Tables S7‐S15.
To ensure our two‐year postperiod specification did not obscure short‐term effects of the transition, we performed a yearly disaggregated postperiod model and an event study to examine the robustness of our findings (Appendix S1: Tables S16‐S21, Figures S2 and S3).
3. RESULTS
3.1. Patient characteristics of those affected by Medicaid Managed Care reshuffling policy
Our initial sample included 432 372 children, ages 5‐18, who were enrolled in Medicaid managed care at some point during the study period (2014‐2016). Restricting the sample to children who were enrolled continuously throughout 2014‐2016 resulted in a sample size of 378 450 children. Of these, 28 537 (8%) had a diagnosis of persistent asthma in 2013. From this patient sample, 21 198 (74%) were plan stayers, and 7339 (26%) were plan switchers. Within plan switchers, 2392 (33%) were PCP stayers, and 4947 (67%) were PCP switchers.
Patients who were forced to switch plans were slightly younger and healthier, more likely to live in rural areas, and had slightly higher baseline levels of ED utilization in 2013 when compared to plan stayers (P < .05 for all comparisons) (Table 1). Among plan switchers, patients who were forced to switch PCPs were less likely to live in rural areas, had slightly higher baseline levels of ED utilization in 2013, and were less likely to be connected to care in 2014 (Table 2).
TABLE 1.
Patient characteristics overall, in base year 2013, and first year of sample 2014, by plan switch status (N = 28 537)
| Patient characteristic | Plan stayer | Plan switcher | P‐value |
|---|---|---|---|
| Male | 0.56 | 0.55 | .31 |
| Mean Age, 2013 | 9.7 | 9.5 | <.001 |
| Van Walraven Score, 2013 | 0.83 | 0.77 | <.001 |
| Rural | 0.37 | 0.39 | .01 |
| Num. of MDs (2015) per 100 000 | 245 | 249 | .16 |
| Num. of Hospital Beds (2014) per 100 000 | 342 | 337 | .14 |
| ED utilization, 2013 | |||
| 0 Visits | 0.58 | 0.55 | <.001 |
| 1 Visit | 0.14 | 0.15 | .05 |
| 2+ Visits | 0.28 | 0.29 | .02 |
| Connected to Care, 2014 | 0.5 | 0.51 | .33 |
| Total | 21 198 | 7339 | |
This table reports patient characteristics by plan switch status. Male indicates the share of male patients in each category. Mean age is the patient's average age in 2013. The Van Walraven Score is the patient's comorbidity score in 2013, scaled such that higher numbers represent patients that are relatively less healthy. Rural indicates the share of patients living in a rural county as defined by Health Resources & Services Administration (HRSA's) Federal Office of Rural Health Policy. Num. of MDs represents the average number of physicians (per 100 000 county residents) as of 2015 available in the county (or counties) the patient lived in. Num. of hospital beds (2014) indicates the average number of hospital beds (per 100 000 county residents) as of 2014 in the county (or counties) the patient lived in. Emergency department is abbreviated as ED, and ED utilization shows the share of patients with 0, 1, or 2 or more visits in 2013. Patients who were connected to care had at least one visit with their assigned PCP in 2014. The total row reports all patients in each Stayer and Switcher category. P values are calculated by comparing Switchers to Stayers.
TABLE 2.
Patient characteristics for plan switchers overall, in base year 2013, and first year of sample 2014, by Primary Care Provider (PCP) Switch Status (N = 7339)
| Patient characteristic | PCP stayer | PCP switcher | P‐value |
|---|---|---|---|
| Male | 0.57 | 0.55 | .09 |
| Mean age, 2013 | 9.4 | 9.5 | .12 |
| Van Walraven Score, 2013 | 0.77 | 0.78 | .53 |
| Rural | 0.41 | 0.38 | .02 |
| Num. MDs (2015) per 100 000 | 249 | 249 | .93 |
| Num. of Hospital Beds (2014) per 100 000 | 340 | 336 | .57 |
| ED utilization, 2013 | |||
| 0 Visits | 0.58 | 0.54 | <.001 |
| 1 Visit | 0.16 | 0.15 | .2 |
| 2+ Visits | 0.26 | 0.31 | <.001 |
| Connected to Care, 2014 | 0.73 | 0.4 | <.001 |
| Total | 2392 | 4947 | |
This table reports patient characteristics by PCP switch status. Male indicates the share of male patients in each category. Mean age is the average age in 2013. The Van Walraven Score is the patient's comorbidity score in 2013, scaled such that higher numbers represent patients that are relatively less healthy. Rural indicates the share of patients living in a rural county as defined by Health Resources & Services Administration (HRSA's) Federal Office of Rural Health Policy. Num. of MDs represents the average number of physicians (per 100 000 county residents) as of 2015 available in the county (or counties) the patient lived in. Num. of hospital beds (2014) indicates the average number of hospital beds (per 100 000 county residents) as of 2014 in the county (or counties) the patient lived in. Emergency department is abbreviated as ED, and ED utilization shows the share of patients with 0, 1, or 2 or more visits in 2013. Patients who were connected to care had at least one visit with their assigned PCP in 2014. The total row reports all patients in each Stayer and Switcher category. P values are calculated by comparing Switchers to Stayers.
3.2. Impact of plan/PCP switch on health care utilization and quality
Table 3 reports the unadjusted trends in the percent of patients with a particular outcome, before and after the January 2015 plan disruption. Outpatient visits, ED visits, and hospitalizations are reported as the percent of patients who had any event in a given quarter. Quality measures are given as the percent of patients who were adherent in a given year. Prior to the plan disruption, 74% of patients had at least one outpatient visit in any given quarter, with less than 10% of patients having an asthma‐specific outpatient visit. Relatively few (15%) patients had an ED visit, and very few (less than 1%) were hospitalized. About one third of patients were adherent to their asthma medication in 2014, and nearly two‐thirds (60%) had an asthma medication ratio of greater than 50%. Following the disruption, nearly all utilization and quality measures decreased, though this decrease cannot necessarily be attributed to the causal effect of the disruption itself.
TABLE 3.
Unadjusted trends in outpatient visits, quality measures, and health care utilization events, before and after January 2015, 2014—2016 (N = 28 537)
| Outcomes | Percent of patients, before Jan. 2015 | Percent of patients, after Jan. 2015 | Percentage‐point difference | 95% CI |
|---|---|---|---|---|
| Outpatient visits | 74 | 70 | −4.05*** | −4.5, −3.6 |
| Asthma outpatient visits | 9.9 | 8.4 | −1.57*** | −1.9, −1.3 |
| Rx adherence | 31 | 25 | −5.7*** | −6.4, −5 |
| Asthma Med. ratio | 61 | 48 | −13*** | −14, −12 |
| ED visits | 15 | 15 | −0.09 | −0.42, −0.24 |
| Asthma ED visits | 1.9 | 1.5 | −0.37*** | −0.49, −0.25 |
| Hospitalization rate | <1% | <1% | 0.02 | −0.07, 0.11 |
| IP asthma | <1% | <1% | −0.11*** | −0.15, −0.07 |
This table shows unadjusted changes in the percent of patients who ever have an outpatient visit, hospitalization, or ED visit in a given quarter, and the percent of patients who were adherent to asthma medication in a given year, before and after the plan switch. The difference column represents the percentage‐point change in the percent of beneficiaries with a given outcome, before versus after the switch. The 95% CI column gives the 95% confidence interval of the difference. Asthma medication adherence is abbreviated as Rx adherence. The asthma medication ratio is abbreviated as Asthma Med Ratio. Hospitalizations for asthma are abbreviated as IP Asthma. Emergency department is abbreviated as ED.
P < .05; **P < .01; ***P < .001.
To estimate the causal effect of the disruption to plan and/or provider, we used a difference‐in‐difference approach. Table 4 reports the regression‐adjusted differences in the change in utilization and quality outcomes after the January 2015 plan disruption, for plan switchers and stayers. We observed that, relative to plan stayers, the number of plan switchers that had at least one general outpatient visit in any given quarter following the switch decreased by an additional 2.06 percentage points [95% CI −2.8, −1.3], from 71% to 66% (compared to plan stayers: 74% to 71%). There was no significant difference between plan switchers and stayers in the use of asthma‐specific outpatient visits (DID −0.4 percentage points, 95% CI −0.89, 0.08), prescription adherence (DID 0.56 percentage points, 95% CI −0.57, 1.7), AMR measures (DID 0.35, 95% CI −1, 1.7), general ED visits (DID −0.12 percentage points, 95% CI −0.72, 0.49), asthma‐specific ED visits (DID −0.06 percentage points, 95% CI −0.3, 0.18), and general hospitalizations (DID −0.08 percentage points, −0.23, 0.08) or asthma‐specific hospitalizations (DID 0.03, 95% CI −0.05, 0.11).
TABLE 4.
Unadjusted trends and regression‐adjusted changes in the percent of patients outpatient visits, quality measures, and health care utilization events, for plan switchers vs stayers before and after January 2015, 2014—2016 (N = 28 537)
| Outcomes | Percent of patients per quarter, before Jan. 2015 | Percent of patients per quarter, after Jan. 2015 | Adjusted percentage‐point difference‐in‐difference | 95% CI |
|---|---|---|---|---|
| Outpatient visits | ||||
| Plan stayer | 74 | 71 | ||
| Plan switcher | 71 | 66 | −2.06*** | −2.82, −1.3 |
| Asthma outpatient visits | ||||
| Plan stayer | 9.8 | 8.3 | ||
| Plan switcher | 10 | 8.4 | −0.4 | −0.89, 0.08 |
| Rx adherence | ||||
| Plan stayer | 32 | 26 | ||
| Plan switcher | 27 | 22 | 0.56 | −0.57, 1.7 |
| Asthma Med ratio | ||||
| Plan stayer | 62 | 49 | ||
| Plan switcher | 58 | 45 | 0.35 | −1, 1.7 |
| ED visits | ||||
| Plan stayer | 15 | 14 | ||
| Plan switcher | 16 | 15 | −0.12 | −0.72, 0.49 |
| Asthma ED visits | ||||
| Plan stayer | 1.7 | 1.4 | ||
| Plan switcher | 2.3 | 1.9 | −0.06 | −0.3, 0.18 |
| Hospitalization rate | ||||
| Plan stayer | <1% | 1 | ||
| Plan switcher | <1% | <1% | −0.08 | −0.23, 0.08 |
| IP asthma | ||||
| Plan stayer | <1% | <1% | ||
| Plan switcher | <1% | <1% | 0.03 | −0.05, 0.11 |
The second and third columns represent the unadjusted percent of patients who ever have an outpatient visit, hospitalization, or ED visit in a given quarter, and the percent of patients who were adherent to asthma medication in a given year. The fourth column shows the adjusted difference‐in‐difference in outcomes, which is given as the percentage‐point change in the percent of patients with an outcome from the preswitch period to the postswitch period. The 95% confidence interval is the range of likely values associated with the treatment effect. Primary care provider is abbreviated as PCP. Asthma medication adherence is abbreviated as Rx adherence. The asthma medication ratio is abbreviated as Asthma Med Ratio. Hospitalizations for asthma are abbreviated as IP Asthma. Emergency department is abbreviated as ED.
P < .05; **P < .01; ***P < .001.
Table 5 reports the regression‐adjusted differences in the change in utilization and quality outcomes after the January 2015 plan disruption, for PCP switchers and stayers among the Plan switchers group. Relative to enrollees who switched plans but not PCPs, there was a smaller decrease in the percent of PCP switchers with at least one asthma‐specific outpatient visit per quarter following the switch. Specifically, the percent of plan switchers with an asthma‐specific outpatient visit in any given quarter following the switch decreased from 9.7% to 8.3%, a relative decrease of 1.6 percentage points [95% CI 0.7, 2.5] less than the approximately 3.4 percentage‐point decrease in the share of PCP stayers (12% to 8.6%) with an asthma‐specific outpatient visit. No other differences between PCP switchers and stayers were significant. Among patients who were connected to care in 2014, the magnitude and direction of the effect of a forced plan and subsequent PCP switch were similar to the primary results. The disaggregated postperiod models, event study, bounding sensitivity analysis, outcome by number of visits, and use of patient controls and AHRF county‐level resources instead of fixed effects returned qualitatively similar results as the primary analysis (Appendix S1: Tables S7‐S21, Figures S2 and S3).
TABLE 5.
Unadjusted trends and regression‐adjusted changes in the percent of patients outpatient visits, quality measures, and health care utilization events, for Primary Care Provider (PCP) switchers vs. stayers (among all plan switchers) before and after January 2015, 2014—2016 (N = 7339)
| Outcomes | Percent of patients per quarter, before Jan. 2015 | Percent of patients per quarter, after Jan. 2015 | Adjusted percentage‐point difference‐in‐difference | 95% CI |
|---|---|---|---|---|
| Outpatient visits | ||||
| PCP stayer | 74 | 68 | ||
| PCP Switcher | 70 | 64 | 0.58 | −0.8, 2 |
| Asthma outpatient visits | ||||
| PCP stayer | 12 | 8.6 | ||
| PCP switcher | 9.7 | 8.3 | 1.6*** | 0.72, 2.5 |
| Rx adherence | ||||
| PCP stayer | 28 | 23 | ||
| PCP switcher | 26 | 21 | 0.57 | −1.5, 2.6 |
| Asthma Med ratio | ||||
| PCP stayer | 61 | 47 | ||
| PCP switcher | 57 | 44 | 0.78 | 1.7, 3.2 |
| ED visits | ||||
| PCP stayer | 14 | 14 | ||
| PCP switcher | 16 | 16 | 0.84 | −0.26, 1.9 |
| Asthma ED visits | ||||
| PCP stayer | 2 | 1.5 | ||
| PCP switcher | 2.5 | 2.1 | 0.09 | −0.33, 0.51 |
| Hospitalization rate | ||||
| PCP stayer | <1% | <1% | ||
| PCP switcher | <1% | <1% | 0.11 | −0.14, 0.37 |
| IP asthma | ||||
| PCP stayer | <1% | <1% | ||
| PCP switcher | <1% | <1% | 0.07 | −0.08, 0.21 |
The second and third columns represent the unadjusted percent of patients who ever have an outpatient visit, hospitalization, or ED visit in a given quarter, and the percent of patients who were adherent to asthma medication in a given year. The fourth column shows the adjusted difference‐in‐difference in outcomes, which is given as the percentage‐point change in the percent of patients with an outcome from the preswitch period to the postswitch period. The 95% confidence interval is the range of likely values associated with the treatment effect. Primary care provider is abbreviated as PCP. Asthma medication adherence is abbreviated as Rx adherence. The asthma medication ratio is abbreviated as Asthma Med Ratio. Hospitalizations for asthma are abbreviated as IP Asthma. Emergency department is abbreviated as ED.
P < .05; **P < .01; ***P < .001.
4. DISCUSSION
In 2015, a large statewide redistribution of a southeastern state's Medicaid managed care enrollees to different plans and PCPs offered a unique natural experiment that helped us to better understand the causal impact of forced plan and PCP discontinuity on patterns of care for children with persistent asthma. Among children who were forced to switch plans, we observed that fewer children had outpatient visits following the switch, relative to children who did not switch plans. Among children who were transitioned to both new plans and providers, we observed a smaller decrease in the percent of children with an asthma‐related outpatient visit following the switch, as compared to children who switched plans but not providers, but that both cohorts experienced a decrease in asthma‐specific outpatient visits following the plan switch. In spite of the decrease in the number of outpatient visits, there was no effect on quality of care measures, or the use of inpatient or ED services.
Current conventional wisdom in health care delivery suggests that disruptions to care represent a threat to patient well‐being. Indeed, some studies have established associational relationships between suboptimal continuity of care and negative outcomes in children with asthma. 32 , 33 , 34 , 35 However, this natural experiment allowed us to more precisely and rigorously parse out the causal effects of disruptions to plan, and (in some cases) provider, continuity on health outcomes of interest. Our findings are supported by results from previous investigations. A study by Ndumele et al found that managed care plan exit was not associated with changes in measures of plan quality in market performance, nor in aggregated measures of patient experience, and Reddy et al found significant adverse effects on patient experience, but not quality, for patients who experienced PCP turnover while enrolled in the Veterans Health Administration. 3 , 36
The lack of association with adverse health utilization and outcomes after the January 2015 disruption has several possible explanations. Increased efforts by the state, plans, and/or new PCPs to encourage disrupted patients to re‐engage in care may have mitigated any consequences of the disruption. Some policy observers have argued that switching plans may have positive implications for patient health, if they are switching away from a low‐quality plan. 37 Furthermore, even among this particularly vulnerable population, a disruption to care—namely health plan and subsequent provider network changes, with no disruption in Medicaid coverage—may not be as threatening to the intermediate‐term health of vulnerable patient cohorts as previously believed.
Our study has a number of limitations. First, we cannot verify that there were not unobserved differences between children who underwent the forced transition and those who did not, but our parallel trends test confirms that prepolicy, the groups were similar on our main outcome measures. Second, while we examined a robust set of outcomes for children with persistent asthma, this cohort may not generalize to children with other chronic conditions. Third, our analysis involved a single state, and thus, our results may not generalize to other state Medicaid managed care programs. However, given the expanding share of managed care within Medicaid, this study sheds light on a question of relevance to a growing number of states. Fourth, a small proportion of the individuals in our sample who switched plans/providers in 2015 may have done so voluntarily, challenging our causal identification strategy; nonetheless, our claims based estimates are largely consistent with state reports indicating the magnitude of the restructuring and our bounding exercise addressed concerns that any observed significant differences in outcomes were a product of baseline levels of switching.
Our study also has significant implications regarding the structuring of Medicaid managed care markets more broadly. We find that two‐thirds of the individuals who involuntarily switched plans also switched primary care providers. This highlights a tension between competition and coordination for Medicaid plans competing in the same markets. Selective provider networks have long been a hallmark of managed care plans; however, if there is little provider overlap between competing plans, it may undermine the very principle of markets that presumes that consumers (enrollees) can move between plans without great friction. Additionally, our study did not observe harmful effects of plan or provider switching for children with persistent asthma. Finding the optimal balance between plan autonomy and patient choice should be a focus of state Medicaid regulators in managed care states.
In an examination of the effect of forced transitions on the use of health services for children with asthma, we did not observe harmful effects of plan or provider switching when continuously enrolled in Medicaid. While the benefits of access to care have been well established, our findings raise questions about the necessity of continuity with a single plan or provider.
Supporting information
Author matrix
Appendix S1
ACKNOWLEDGMENTS
Joint Acknowledgment/Disclosure Statement: This publication was made possible by the Yale CTSA grant UL1TR000142 from the National Center for Advancing Translational Science (NCATS), NIH.
Financial Disclosures and Conflict of Interest: In the past 36 months, Dr. Ross received research support through Yale University from Medtronic, Inc. and the Food and Drug Administration (FDA) to develop methods for postmarket surveillance of medical devices (U01FD004585), from the Centers of Medicare and Medicaid Services (CMS) to develop and maintain performance measures that are used for public reporting (HHSM‐500‐2013‐13018I), and from the Blue Cross Blue Shield Association to better understand medical technology evaluation; Dr. Ross currently receives research support through Yale University from Johnson and Johnson to develop methods of clinical trial data sharing, from the Food and Drug Administration to establish Yale‐Mayo Clinic Center for Excellence in Regulatory Science and Innovation (CERSI) program (U01FD005938), from the Medical Device Innovation Consortium as part of the National Evaluation System for Health Technology (NEST), from the Agency for Healthcare Research and Quality (R01HS022882), from the National Heart, Lung and Blood Institute of the National Institutes of Health (NIH) (R01HS025164, R01HL144644), and from the Laura and John Arnold Foundation to establish the Good Pharma Scorecard at Bioethics International and to establish the Collaboration for Research Integrity and Transparency (CRIT) at Yale. The other authors have no financial disclosures or conflicts of interest to disclose.
Piwnica‐Worms K, Staiger B, Ross JS, Rosenthal MS, Ndumele CD. Effects of forced disruption in Medicaid managed care on children with asthma. Health Serv Res. 2021;56:668–676. 10.1111/1475-6773.13643
REFERENCES
- 1. Total Medicaid Managed Care Enrollment. 2016. https://www.kff.org/medicaid/state‐indicator/total‐medicaid‐mc‐enrollment/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D
- 2. Schneider A. How can we tell whether Medicaid MCOs are doing a good job for kids? 2018. https://ccf.georgetown.edu/2018/02/26/how‐can‐we‐tell‐whether‐medicaid‐mcos‐are‐doing‐a‐good‐job‐for‐kids/
- 3. Ndumele CD, Schpero WL, Schlesinger MJ, Trivedi AN. Association between health plan exit from Medicaid Managed Care and quality of care, 2006–2014. JAMA. 2017;317:2524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Landon BE, Wilson IB, Cleary PD. A conceptual model of the effects of health care organizations on the quality of medical care. JAMA. 1998;279:1377‐1382. [DOI] [PubMed] [Google Scholar]
- 5. McCue MJ, Bailit MH. Assessing the financial health of Medicaid managed care plans and the quality of patient care they provide. Issue Brief Commonw Fund. 2011;11:1‐16. [PubMed] [Google Scholar]
- 6. Garfield R, Hinton E, Cornachione E, Hall C. Medicaid managed care plans and access to care: results from the Kaiser Family Foundation 2017 Survey of Medicaid Managed Care Plans. 2018. https://www.kff.org/report‐section/medicaid‐managed‐care‐plans‐and‐access‐to‐care‐plan‐characteristics‐enrollees‐and‐services/
- 7. Chernew ME, Wodchis WP, Scanlon DP, McLaughlin CG. Overlap in HMO physician networks. Health Aff Proj Hope. 2004;23:91‐101. [DOI] [PubMed] [Google Scholar]
- 8. Coleman AT, Teach SJ, Sheehan WJ. Inner‐City asthma in childhood. Immunol Allergy Clin North Am. 2019;39:259‐270. [DOI] [PubMed] [Google Scholar]
- 9. Haggerty JL, Reid RJ, Freeman GK, Starfield BH, Adair CE, McKendry R. Continuity of care: a multidisciplinary review. BMJ. 2003;327:1219‐1221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Barnett ML, Song Z, Rose S, Bitton A, Chernew ME, Landon BE. Insurance transitions and changes in physician and Emergency Department Utilization: an observational study. J Gen Intern Med. 2017;32:1146‐1155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Lavarreda SA, Gatchell M, Ponce N, Brown ER, Chia YJ. Switching health insurance and its effects on access to Physician Services. Med Care. 2008;46:1055‐1063. [DOI] [PubMed] [Google Scholar]
- 12. Bindman AB, Chattopadhyay A, Auerback GM. Medicaid re‐enrollment policies and children's risk of hospitalizations for ambulatory care sensitive conditions. Med Care. 2008;46:1049‐1054. [DOI] [PubMed] [Google Scholar]
- 13. NCQA health plan face sheet. https://www.ncqa.org/hedis/.
- 14. Wakefield DB, Cloutier MM. Modifications to HEDIS and CSTE algorithms improve case recognition of pediatric asthma. Pediatr Pulmonol. 2006;41:962‐971. [DOI] [PubMed] [Google Scholar]
- 15. Core Set of Children's Health Care Quality Measures for Medicaid and CHIP: Technical specifications and resource manual for federal fiscal year 2017 reporting. 2017.
- 16. van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser Comorbidity measures into a point system for hospital death using administrative data. Med Care. 2009;47:626‐633. [DOI] [PubMed] [Google Scholar]
- 17. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8‐27. [DOI] [PubMed] [Google Scholar]
- 18. Johnston EE, Alvarez E, Saynina O, Sanders L, Bhatia S, Chamberlain LJ. End‐of‐life intensity for adolescents and young adults with cancer: a Californian Population‐Based Study that shows disparities. J Oncol Pract. 2017;13:e770‐e781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Tejwani R, Jiang R, Wolf S, et al. Contemporary demographic, treatment, and geographic distribution patterns for disorders of sex development. Clin Pediatr. 2018;57:311‐318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Wang H‐HS, Tejwani R, Wolf S, Wiener JS, Routh JC. Readmissions, unplanned emergency room visits, and surgical retreatment rates after anti‐reflux procedures. J Pediatr Urol. 2017;13:507.e1‐507.e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Zein JG, Udeh BL, Teague WG, et al. Impact of age and sex on outcomes and hospital cost of acute asthma in the United States, 2011–2012. PLoS One. 2016;11:e0157301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Ronksley PE, Kobewka DM, McKay JA, Rothwell DM, Mulpuru S, Forster AJ. Clinical characteristics and preventable acute care spending among a high cost inpatient population. BMC Health Serv. Res. 2016;16:165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat. 1979;6:65‐70. [Google Scholar]
- 24. Heymann PW, Carper HT, Murphy DD, et al. Viral infections in relation to age, atopy, and season of admission among children hospitalized for wheezing. J Allergy Clin Immunol. 2004;114:239‐247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Kimes D, Levine E, Timmins S, Weiss SR, Bollinger ME, Blaisdell C. Temporal dynamics of emergency department and hospital admissions of pediatric asthmatics. Environ Res. 2004;94:7‐17. [DOI] [PubMed] [Google Scholar]
- 26. Heymann PW, Zambrano JC, Rakes GP. Virus‐induced wheezing in children. Immunol Allergy Clin North Am. 1998;18:35‐47. [Google Scholar]
- 27. Wisniewski JA, McLaughlin AP, Stenger PJ, et al. A comparison of seasonal trends in asthma exacerbations among children from geographic regions with different climates. Allergy Asthma Proc. 2016;37:475‐481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Valet RS, Gebretsadik T, Carroll KN, et al. High asthma prevalence and increased morbidity among rural children in a Medicaid cohort. Ann Allergy Asthma Immunol. 2011;106:467‐473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Baltrus P, Xu J, Immergluck L, et al. Individual and county level predictors of asthma related emergency department visits among children on Medicaid: a multilevel approach. J Asthma. 2017;54:53‐61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Malhotra K, Baltrus P, Zhang S, McRoy L, Immergluck LC, Rust G. Geographic and racial variation in asthma prevalence and emergency department use among Medicaid‐enrolled children in 14 southern states. J Asthma. 2014;51:913‐921. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Dombkowski KJ, Cabana MD, Cohn LM, Gebremariam A, Clark SJ. Geographic variation of asthma quality measures within and between health plans. Am J Manag Care. 2005;11:765‐772. [PubMed] [Google Scholar]
- 32. Christakis DA, Mell L, Koepsell TD, Zimmerman FJ, Connell FA. Association of lower continuity of care with greater risk of emergency department use and hospitalization in children. Pediatrics. 2001;107:524‐529. [DOI] [PubMed] [Google Scholar]
- 33. Utidjian LH, Fiks AG, Localio AR, et al. Pediatric asthma hospitalizations among urban minority children and the continuity of primary care. J Asthma. 2017;54:1051‐1058. [DOI] [PubMed] [Google Scholar]
- 34. Cree M, Bell NR, Johnson D, Carriere KC. Increased continuity of care associated with decreased hospital care and emergency department visits for patients with asthma. Dis Manag. 2006;9:63‐71. [DOI] [PubMed] [Google Scholar]
- 35. Huang S‐T, Wu S‐C, Hung Y‐N, Lin I‐P. Effects of continuity of care on emergency department utilization in children with asthma. Am J Manag Care. 2016;22:e31‐e37. [PubMed] [Google Scholar]
- 36. Reddy A, Pollack CE, Asch DA, Canamucio A, Werner RM. The effect of primary care provider turnover on patient experience of care and ambulatory quality of care. JAMA Intern Med. 2015;175:1157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Long SK, Yemane A. Commercial plans in Medicaid managed care: understanding who stays and who leaves. Health Aff. 2005;24:1084‐1094. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Author matrix
Appendix S1