This NQF-endorsed measure (NQF#2797), Transcranial Doppler Ultrasonography (TCD) Screening among Children with Sickle Cell Anemia (SCA), establishes a claims-based method for identifying receipt of TCD screening among populations of children with sickle cell anemia. The measure was created by the Quality Measurement, Evaluation, Testing, Review, and Implementation Consortium (QMETRIC) operating through the University of Michigan, a Pediatric Quality Measures Program grantee. This measure was initially endorsed by the National Quality Forum in May 2016 and has maintained endorsement since that time. This measure was also recommended for inclusion in the 2016 and 2018 Medicaid and CHIP Child Core Set (Child Core Set) by the Measure Applications Partnership.
Specifically, the measure calculates the percentage of children ages 2 through 15 years old with SCA (Hemoglobin [Hb] SS or HbSβ0-thalassemia) who received at least one TCD screening within the measurement year. A higher proportion indicates better performance as reflected by appropriate testing. The measure specifications are reflective of the guidelines from the National Heart, Lung, and Blood Institute (NHLBI), as well as recent guidance from the American Society for Hematology (ASH) that was updated after development of this measure, and the performance scores calculated through this measure identify areas in need of improvement.
The following text was taken from the NQF Measure Submission (2016): Children with SCA have over three hundred times the stroke risk than children with normal hemoglobin (Verduzco and Nathan, 2009). Without intervention, approximately 11% of children with SCA will have a stroke by age 20 (Verduzco and Nathan, 2009; Ohene-Frempong et al., 1998). TCD ultrasonography measures the blood velocities within the cerebral vessels (Adams et al., 1997; Adams et al., 1992). Children over the age of 2 with a time-average mean maximum blood flow velocity of 200cm/sec or greater as measured by TCD ultrasonography have been shown to have 27 times the risk of stroke than children with velocities less than 200cm/sec. This corresponds to a 40% risk of stroke among those with high velocities within 3 years (Adams et al., 1997). Initiation of chronic blood transfusions reduces the risk of stroke by 92% among children at highest risk of stroke as identified through TCD screening (Adams et al., 1997; Adams et al., 1992). TCD screening is a reasonable method to assess stroke risk among children with SCA, as it is safe, non-invasive and low cost (Markus, 2000). Although other predictors of stroke have been examined, such as hematocrit levels and white blood cell count, TCD velocities have been shown to be the only independent predictor of stroke (Adams et al., 1992). Given the importance of TCD screening to stroke prevention among children with SCA, the National Heart, Lung, and Blood Institute (NHLBI) recommends each child receive one TCD screen per year from ages 2 through 15 years (National Heart, Lung, and Blood Institute, 2014). Although the benefits of TCD screening among children with SCA have been known since the late nineties, prior studies indicate that TCD screening rates are low. However, these reports are limited in their generalizability, as they are often focused on a single healthcare provider or registry. This measure establishes a claims-based method for identifying receipt of TCD screening among larger and broader populations of children with SCA. The measure specifications are reflective of the guidelines from the NHLBI, and the performance scores calculated through this measure will identify areas in need of improvement in receipt of TCD screening among children with SCA.
Intended End Users
It is intended that this measure will be used at the state and health plan levels. Although the measure is valid at the health system level, some health systems may not be able to implement the measure as it may not be possible for them to determine if TCD screenings were obtained outside of their health system. The measure may be used at the state-level for public reporting, public health/disease surveillance, and quality improvement efforts within the state and among Medicaid health plans. Health plans may use the measure for public reporting and quality improvement efforts within the organization.
Measure Implementation Feasibility at State, Health Plan and Health System Levels
This TCD measure may be implemented at the state, health plan and health system levels, although as noted above, health systems may find it more challenging to implement. To specifically address attribution within the context of implementing the measure at different levels, the QMETRIC team assessed the ability to fold the measure up and down using aggregated and disaggregated data. Identifying accurate denominator populations at the state, health plan, and health system levels was the focus of the team for folding down the measure. Efforts to fold the measure up required ensuring measure validity at each level.
Folding Down from State Level
To fold the measure down from the state to the health plan and from the state to the health system level, the QMETRIC team sought to determine which one health plan and one health system was accountable for the care of each child with SCA. The approaches the QMETRIC team explored are detailed below.
Health Plan Attribution
QMETRIC first identified all children who were enrolled in Michigan Medicaid for a calendar year, regardless of which Medicaid health plan (MHP) the child was enrolled in. A full year of enrollment (12 months continuous enrollment) is necessary to calculate this measure. Guidelines from the NHLBI strongly recommend each child with SCA should receive one TCD screen within a 12-month period; therefore, this time period is necessary to capture guideline-adherent quality of care. To attribute children to MHPs, the QMETRIC team attributed each person-month of enrollment to an MHP and summed months across the measurement year. They then assessed the number of children who were attributed to each MHP when requiring seven months, 11 months, and 12 months enrollment in a calendar year. For each continuous enrollment requirement, the proportion of children with SCA who could be attributed to an MHP was calculated. Likewise, the proportion of children with SCA who were enrolled in Michigan Medicaid for a calendar year but not attributable to an MHP was calculated for each of the enrollment requirements. Performance scores were calculated for each MHP based on the enrollment requirements.
Health System Attribution
To attribute children to a health system, the QMETRIC team conducted a comprehensive assessment of other attribution models. A series of standards to determine health system attribution for children with SCA based on patterns of care was developed. First, all children with SCA enrolled in Michigan Medicaid for a 12-month period were identified. Secondly, each child’s health services encounters, including inpatient admissions, outpatient visits, and emergency department visits, were characterized and summed. Then the QMETRIC team identified the health system that provided care during each encounter, using the billing provider information (i.e., the National Provider Identifier (NPI)) and a manual review to match the NPI to the correct health system. Next, for each child, the percentage of their encounters at each health system was calculated. The QMETRIC team then applied a series of “cut-offs” for attribution to a health system. Specifically, a child was attributed to a health system if 90 percent, 75 percent, or 51 percent of the child’s encounters occurred within that hospital system. Based on the above encounter requirements, the proportion of children with SCA that could be attributed to a specific health system was calculated.
Folding Up from Health Plan Level
To assess if a measure could be folded up from the health plan to the state level, it was necessary to determine if the measure was valid at the health plan level. To validate the measure at the health plan level, the QMETRIC team compared two sets of performance scores: (1) rates calculated through attributing children using state-level Medicaid data to specific MHPs (process described above), and (2) rates calculated directly by each of the three partner MHPs using their in-house data.
Folding Down from State Level
Health Plan Attribution
The following results used 2018 Michigan Medicaid data to illustrate the impact of the varying eligibility requirements within a single MHP (seven months, 11 months, and 12 months). Among children continuously enrolled in Michigan Medicaid for the calendar year, the proportion of children that could not be attributed to a specific MHP varied significantly across the years 2010-2018. Across the entire study period, the number of children enrolled in fee-for-service for the entire year decreased substantially. The QMETRIC team was able to calculate performance scores for both the attributed and the non-attributed children, providing data to illustrate opportunities for improvement across all children enrolled in Michigan Medicaid.
Health System Attribution
The following results used 2017 Michigan Medicaid data to illustrate the impact of varying encounter requirements for a health system. To assess the validity of each approach, QMETRIC expanded the age ranges to children with SCA from one year to 17 years specifically for understanding the tradeoffs of health system attribution cutoffs; no changes were made to age criteria for calculation of the TCD measure. During this time, there were 402 children with 7,131 encounters. Among these children, 280 children (70 percent) received at least 90 percent of their care at the same health system. An additional 44 children (total of 81 percent) received at least 75 or 51 percent of their care at the same health system.
Folding Up from Health Plan Level
There were no significant differences in performance scores for MHPs when comparing 2017 and 2018 performance scores obtained by attributing children from the state level to three MHPs and performance scores obtained directly from these MHPs.
The QMETRIC findings regarding the attribution of children to one MHP have implications for all pediatric quality measures. Other measures will face similar challenges, i.e., there will be children who cannot be attributed to a single MHP for the measure but are enrolled within a state Medicaid plan for the enrollment period. This impacts the targets of QI initiatives to improve the performance scores of the measure. Development of sustainable strategies at the state Medicaid level that incentivize MHPs to retain pediatric members may be one such opportunity. This would allow MHPs to identify members to target for improvement in performance scores. Otherwise, mechanisms to assign responsibility for the quality of care of these children should be in place. Care should be taken to ensure that children who enroll in different health plans across one year do not fall through the cracks.
The findings regarding the attribution of children to one health system also have implications for quality measurement. As discussed above, there are no validated methods for health system attribution. The QMETRIC method, which assessed several cutoffs and assessed the attribution tradeoffs at each cutoff, provided evidence to support the decision to use the 75 percent of encounters at one health system as the attribution method. Other chronic conditions may have different patterns of care. Therefore, health system attribution models should be based on both a conceptual framework of how children access the healthcare system, as well as an evidence-based approach as described above, to perform the most appropriate attribution.
Based on the PQMP QMETRIC findings, this measure can be folded down from the state level to the health plan level and from the state level to the health system level. Folding down the measures from the state to health system level allowed the QMETRIC team to identify performance gaps, enabling quality improvement initiatives to be implemented at the health system level.
This measure can be folded up from the health plan to the state level. However, all measures must be assessed for validity at each level prior to attempting to fold up. Further, it is important to consider which children may not be included when folding up. For example, as described above, there is a significant proportion of children who would not be included in the measure when folding up from the health plan to the health system, as these children may not have been enrolled in the same MHP for the year.
Adams RJ, McKie VC, Carl EM, et al. Long-term stroke risk in children with sickle cell disease screened with transcranial Doppler. Ann Neurol Nov 1997;42(5):699-704.
Adams R, McKie V, Nichols F, et al. The use of transcranial ultrasonography to predict stroke in sickle cell disease. N Engl J Med Feb 27 1992;326(9):605-610.
Markus HS. Transcranial Doppler ultrasound. Br Med Bull 2000;56(2):378-388.
National Heart Lung and Blood Institute. Evidence Based Management of Sickle Cell Disease. 2014; https://www.nhlbi.nih.gov/health-topics/evidence-based-management-sickle-cell-disease. Accessed August 10, 2021.
Ohene-Frempong K, Weiner SJ, Sleeper LA, et al. Cerebrovascular accidents in sickle cell disease: rates and risk factors. Blood Jan 1 1998;91(1):288-294.
Verduzco LA, Nathan DG. Sickle cell disease and stroke. Blood 2009;114(25):5117-5125. doi:10.1182/blood-2009-05-220921