Big Data, Little Data, and Care Coordination for Medicare Beneficiaries with Medigap Coverage

Abstract

Most healthcare data warehouses include big data such as health plan, medical, and pharmacy claims information for many thousands and sometimes millions of insured individuals. This makes it possible to identify those with multiple chronic conditions who may benefit from participation in care coordination programs meant to improve their health. The objective of this article is to describe how large databases, including individual and claims data, and other, smaller types of data from surveys and personal interviews, are used to support a care coordination program. The program described in this study was implemented for adults who are generally 65 years of age or older and have an AARP(®) Medicare Supplement Insurance Plan (i.e., a Medigap plan) insured by UnitedHealthcare Insurance Company (or, for New York residents, UnitedHealthcare Insurance Company of New York). Individual and claims data were used first to calculate risk scores that were then utilized to identify the majority of individuals who were qualified for program participation. For efficient use of time and resources, propensity to succeed modeling was used to prioritize referrals based upon their predicted probabilities of (1) engaging in the care coordination program, (2) saving money once engaged, and (3) receiving higher quality of care. To date, program evaluations have reported positive returns on investment and improved quality of healthcare among program participants. In conclusion, the use of data sources big and small can help guide program operations and determine if care coordination programs are working to help older adults live healthier lives.

Big data analytics in healthcare: a systematic literature review

“…Applications of BDA in healthcare are gradually increasing with the growing volume of big data in this context (Galetsi and Katsaliaki 2019;Kamble et al 2019). Among the possible sources of big data in healthcare are heterogeneous and multi-spectral observations, such as patient demographics (Malik, Abdallah, and Ala'raj 2018), treatment history (Ozminkowski et al 2015), and diagnostic reports (Amirian et al 2017). Mehta and Pandit (2018) suggest that such data may be structured (e.g., genotype, phenotype, or genomics data) or unstructured (e.g., clinical notes, prescriptions, or medical imaging).…”

“…Summary of findings of reviewed studies. (Chandola, Sukumar, and Schryver 2013); integrating techniques (Cheng, Kuo, and Zhou 2018); insights from sociodemographic data (Amirian et al 2017;Narayanan and Greco 2016); insurance claims (Chandola, Sukumar, and Schryver 2013); fraud identification (Chandola, Sukumar, and Schryver 2013); structural degradation modelling (Chehade and Liu 2019); macro-level phenomena (Cheng, Kuo, and Zhou 2018); public-health policy (Christensen et al 2018); social welfare policies (Wu et al 2016) Disease prediction Serious medical conditions (Chen et al 2017;Hadi et al 2019;Yasin and Rao 2018); gestational diabetes mellitus (Moreira et al 2018); diabetes (George, Chacko, and Kurien 2019;Gowsalya, Krushitha, and Valliyammai 2014); disease patterns (De Silva et al 2015); efficient risk profiling (Lin et al 2017); diagnostic frameworks (Babar et al 2016); prediction models (Manogaran et al 2018); prioritising individuals (Ozminkowski et al 2015;Sabharwal, Gupta, and Thirunavukkarasu 2016); personalised healthcare apps (Tseng et al 2017); patient monitoring (Christensen et al 2018;Sabharwal, Gupta, and Thirunavukkarasu 2016); disease-based monitoring systems (Bravo et al 2018); real-time assessment in m-Health (Bravo et al 2018); secure living environment for elderly (Jin et al, 2016) Strategy formulation BDA-based capabilities (Austin and Kusumoto 2016); investment in BDA (Sabharwal, Gupta, and Thirunavukkarasu 2016); efficient resource allocation (Gowsalya, Krushitha, and Valliyammai 2014); knowledge management …”

“…Medical records act as building blocks of historical data in healthcare, particularly for sourcing patient data. These are commonly obtained from diagnostic reports (Amirian et al 2017;De Silva et al 2015), hospital registers (Austin and Kusumoto 2016;Babar et al 2016), and patients' history (Ozminkowski et al 2015;Zaragoza, Kim, and Chung 2017). Medical records are often available in electronic formats, such as electronic medical records containing a patient's treatment history from past visits to a doctor.…”

“…Contexts of reviewed studies Lin et al 2017)(Chandola, Sukumar, and Schryver 2013;Ozminkowski et al 2015); hearing aids(Christensen et al 2018); living environment(Jin et al, 2016); sports behaviour(Tseng et al 2017); wearable devices(Li et al 2015;Wu et al 2016; …”

Quality of care assessment for people with multimorbidity

“…The search retrieved 435 documents after removal of duplicates (Fig. ), and after eligibility screening, a total of 87 documents were finally included (Appendix ).…”