Abstract
Context:
Populations with complex chronic conditions (CCCs), especially those reliant on medical technology, face disproportionate risks during disasters. Medicaid enrollees make up a large portion of these vulnerable populations, yet states often lack rapid identification systems to inform disaster planning.
Program:
Florida Medicaid developed a claims-based method to identify enrollees with CCCs, including those dependent on medical technology, to support emergency preparedness and response activities.
Implementation:
Using the Florida Medicaid Management Information System, the team applied a validated diagnostic and procedural code framework to classify enrollees into 12 CCC categories. Data were analyzed by age, geography, and technology dependence. Health plans received real-time reports before and after hurricanes in 2022 to 2024.
Evaluation:
Among 4.4 million enrollees, 7.2% had CCCs, and 18.2% of those were technology dependent. Geographic mapping showed higher concentrations in rural and coastal areas prone to disasters. Health plans reported using the data to contact members and coordinate services during hurricane recovery.
Discussion:
This effort demonstrates the feasibility and utility of applying claims data to support disaster management. Other states may adapt this approach to improve emergency response and continuity of care for Medicaid’s most medically vulnerable populations.
Keywords: chronic conditions, disaster preparedness, health policy, Medicaid, population health
Introduction
Complex chronic conditions (CCCs) are persistent, multisystem health conditions requiring long-term management and specialized care, and include advanced heart disease, kidney failure, and technology-dependent conditions such as ventilator use.1 Individuals with CCCs are particularly vulnerable during disasters because they depend on uninterrupted access to medical care, power-dependent equipment, and medications.2 Medicaid serves a high proportion of people with CCCs, many of whom face additional challenges such as poverty, transportation barriers, and limited access to specialty care.3 These overlapping barriers can reduce resilience during disasters, particularly in areas where health infrastructure is already under strain. Together, these factors elevate the risk of serious health consequences during emergencies, as documented after events such as Hurricanes Katrina (2005), Sandy (2012), and Irma (2017), when disruptions in health care services led to increased hospitalizations and mortality among medically fragile populations.4-6
Florida’s frequent hurricane seasons have emphasized the need for targeted disaster preparedness for Medicaid enrollees with CCCs. However, state agencies and health plans have lacked a standardized, scalable method to identify and map this population across regions.7 Without timely, actionable data, emergency management and health care providers struggle to allocate resources, plan for power-dependent medical needs, and maintain continuity of care during and after disasters.8
To address this gap, Florida Medicaid adapted a claims-based classification system to identify individuals with CCCs using administrative claims data.9 This methodology, applied to the full Medicaid population, was designed to support state agencies and their contracted health plans in identifying high-risk enrollees before hurricane seasons. This Practice Full Report describes the development, implementation, and application of this system to support disaster planning and emergency response efforts for Florida’s Medicaid population.
Methods
Program setting and data sources
In 2023-2024, Florida Medicaid provided health coverage to over 4 million low-income individuals, including children, pregnant women, adults with disabilities, and older adults. To support emergency preparedness for these vulnerable populations, Florida Medicaid utilized administrative data from the Florida Medicaid Management Information System (FMMIS). This system contained enrollment records and paid claims, including diagnosis and procedure codes, for all Medicaid enrollees statewide. Demographic data such as age, sex, race or ethnicity, and county of residence were obtained for Medicaid recipients with health plan enrollment and claims data from July 1, 2023, to June 30, 2024.
Development of the claims-based classification method
To identify Medicaid enrollees with CCCs, the classification system developed by Feudtner (2014) was adapted by translating the original SAS-based algorithm into structured query language for integration into the FMMIS.9 This adaptation required restructuring diagnostic and procedural logic to operate within FMMIS’s relational database architecture, including the use of standardized ICD-10-CM, Current Procedural Terminology, and Healthcare Common Procedure Coding System code sets. The structured query language queries were designed to apply temporal logic (identifying enrollees with at least 2 qualifying claims separated by 30 days or more) and to enable population-wide execution across multiple years of claims data, ensuring scalability, reproducibility, and compatibility with FMMIS performance constraints.
The original classification system by Feudtner (2014) was developed and validated in pediatric populations to identify children with life-limiting and multisystem conditions requiring specialized care.9 The classification system used ICD-9 and ICD-10 diagnosis codes to identify medical conditions expected to last at least 12 months (unless death intervened) and involve either several different organ systems or 1 organ system severely enough to require specialty pediatric care and likely hospitalization in a tertiary care center. Conditions were grouped into mutually exclusive categories such as neuromuscular, cardiovascular, respiratory, renal, gastrointestinal, hematologic or immunologic, metabolic, congenital or genetic, malignancy, and neonatal disorders. The original system was developed using large national pediatric inpatient datasets to assess mortality risk and resource needs among hospitalized children.
While the initial validation focused on children, the underlying logic, categorizing conditions based on diagnostic codes that indicate severity, chronicity, and complexity is broadly applicable to other high-risk populations served by Medicaid. Thus, this case-finding algorithm was applied across the entire Medicaid enrollee population, which, while predominantly pediatric, also included impoverished pregnant women, adults with disabilities, and older adults. These groups are similarly affected by chronic, multisystem diseases and are often reliant on continuous, resource-intensive care. Also, these populations can experience overlapping vulnerabilities and rely on medical technologies such as dialysis, ventilators, or feeding tubes, much like medically complex children. Additionally, applying the algorithm across this broader population helps capture not only clinical complexity but also compounding structural vulnerabilities such as poverty, transportation, and disrupted access to care that intensify health risks during service disruptions. By extending the classification system to all enrollees and applying consistent temporal and coding criteria, the algorithm provided a standardized, scalable approach to identifying individuals across the lifespan who are at highest risk of poor outcomes during health care disruptions.
Technology-dependent enrollees were identified using the CCC classification system’s Technology Dependence category, which includes external, life-sustaining equipment such as mechanical ventilators, feeding tubes, dialysis machines, and central lines, but excludes internal implantable devices like pacemakers and insulin pumps.
Calculation of prevalence and geographic distribution
For each identified enrollee, the CCC category or categories were recorded along with their demographic and geographic information. Enrollees with codes indicating technology dependence (eg, ventilators, feeding tubes, dialysis) were flagged as a subgroup of particular interest for emergency response planning.
The prevalence of CCCs was calculated at both the state and county levels. To account for differences in age distributions across counties, age-adjusted prevalence rates were calculated using the direct standardization method with the statewide Medicaid population as the reference.10 Age-groups used for standardization included: <1, 1 to 5, 6 to 13, 14 to 20, 21 to 34, 35 to 44, 45 to 54, 55 to 64, 65 to 74, 75 to 84, and 85+ years.
Application in emergency preparedness
Lists of enrollees identified as having technology-dependent CCCs were shared with Florida Medicaid health plans prior to and following hurricanes making landfall in Florida in 2022 to 2024. Health plans used these lists to support outreach, care coordination, and resource deployment in regions affected by hurricanes and associated power outages. County-level prevalence maps and summary reports were also generated to support state and local emergency management efforts.
Results
Prevalence of CCCs in the Florida Medicaid population
As of June 2024, Florida Medicaid covered 4 428 516 individuals (Table 1). Using the adapted claims-based classification method, 317 592 individuals (7.2%) were identified as having at least 1 CCC. Among these, 57 725 (18.2%) were classified as technology dependent, representing 1.3% of the total Medicaid population.
TABLE 1.
Demographic Characteristics of Florida Medicaid Enrollees With Complex Chronic Conditions (CCCs) and Technology Dependence, 2024
| Characteristic | Total Medicaid Enrollees | Enrollees With CCC, n (%) | CCC Enrollees With Technology Dependence, n (%) |
|---|---|---|---|
| Total | 4 428 516 | 317 592 (7.2) | 57 725 (18.2% of CCC) |
| Sex | |||
| Female | 2 596 467 | 189 330 (60) | 29 401 (15.5) |
| Male | 1 832 049 | 128 262 (40) | 28 324 (22.1) |
| Age-Group | |||
| 2 064 371 | 41 674 (13) | 9519 (22.8) | |
| 18-44 | 1 170 446 | 61 608 (19) | 9645 (15.7) |
| 45-64 | 433 515 | 94 064 (30) | 16 640 (17.7) |
| 65+ | 760 184 | 120 246 (38) | 21 921 (18.2) |
| Race/Ethnicity | |||
| White | 158 593 | 98 390 (31) | 17 091 (17.4) |
| Black/African American | 1 096 494 | 56 364 (18) | 11 519 (20.4) |
| Hispanic | 1 167 611 | 65 238 (21) | 9053 (13.9) |
| Other/Unknown | 2 005 818 | 97 600 (30) | 20 062 (20.6) |
Demographic analysis showed that CCCs were identified across all age-groups, with the largest proportion in adults aged 65 and older (38% of all CCC cases), followed by adults aged 45 to 64 years (30%). Pediatric enrollees under 18 years of age accounted for 13% of all CCC cases but had the highest percentage of technology dependence at 22.8%. Among adults, technology dependence was most common in the 45 to 64 age-group (17.7%) and the 65 and older group (18.2%).
Sex-specific analyses revealed that while female enrollees accounted for a greater share of the total CCC population (60%), male enrollees had a higher proportion of technology dependence (22.1% compared to 15.5% among females). Racial and ethnic breakdowns showed that Black enrollees had a higher rate of technology dependence (20.4%) compared to White (17.4%) and Hispanic (13.9%) enrollees.
Distribution by condition type and technology dependence
Among the 12 CCC categories, metabolic conditions (ie, diabetes, obesity, metabolic syndrome) were the most common, affecting 220 682 individuals (69.5% of all CCC cases) (Table 2). Cardiovascular (29.7%) and gastrointestinal (18.1%) conditions were also prevalent. The highest rates of technology dependence were observed in individuals with transplantation (99.8%), respiratory (76.2%), renal (74.5%), and gastrointestinal (42.9%) conditions.
TABLE 2.
Categories of Complex Chronic Conditions and Rates of Technology Dependence Among Florida Medicaid Enrollees, 2024
| CCC Category | Total With CCC, n (%) | CCC Enrollees With Technology Dependence, n (%) |
|---|---|---|
| Metabolic | 220 682 (69.5) | 31 367 (14.2) |
| Cardiovascular | 94 354 (29.7) | 32 877 (34.8) |
| Gastrointestinal | 57 382 (18.1) | 24 629 (42.9) |
| Neurologic/Neuromuscular | 41 192 (13.0) | 12 063 (29.3) |
| Congenital/Genetic | 33 882 (10.7) | 6811 (20.1) |
| Renal | 21 997 (6.9) | 16 383 (74.5) |
| Hematologic/Immunologic | 19 217 (6.1) | 2098 (10.9) |
| Respiratory | 8680 (2.7) | 6612 (76.2) |
| Neonatal | 4819 (1.5) | 2063 (42.8) |
| Malignancy | 1429 (0.5) | 329 (23.0) |
| Transplantation | 1350 (0.4) | 1347 (99.8) |
Of the 317 592 individuals identified with CCCs, nearly half (44.8%) had 2 or more CCC categories documented.
Geographic distribution of CCCs
Age-adjusted prevalence rates of CCCs varied across Florida’s 67 counties, ranging from 5.7 to 13.2 per 100 000 Medicaid enrollees (Figure 1). Higher prevalence rates were observed in rural and coastal regions, including counties in the Panhandle, Big Bend, and North Central Florida. The percentage of technology dependence among CCC enrollees also varied by county, ranging from 11.4% to 24.7%, with the highest concentrations in North Florida and coastal regions such as Pinellas County (Figure 2).
FIGURE 1.
Age-Adjusted County-Level Prevalence of Complex Chronic Conditions Among Florida Medicaid Enrollees, 2024. Heatmap of CCC prevalence per 100 000 Medicaid enrollees by county.
FIGURE 2.
Percentage of Technology Dependence Among CCC Enrollees by County, 2024. Heatmap showing the proportion of technology-dependent enrollees among CCCs by County.
Application in hurricane preparedness
During the 2022 to 2024 hurricane seasons, Florida Medicaid operationalized the CCC case finding method by generating lists of enrollees identified as technology dependent. In particular, these lists were provided to Medicaid health plans serving regions affected by 2 of the stronger hurricanes, Hurricanes Ian (2022) and Idalia (2023), enabling targeted outreach to ensure continuity of care for medically vulnerable enrollees. Health plans reported using these data to proactively contact members with CCCs, prioritize wellness checks for those identified as technology dependent, and coordinate delivery or replacement of critical medical equipment such as ventilators, oxygen supplies, and feeding pumps. In several cases, plans arranged temporary relocation support for medically fragile enrollees whose homes lost power or became uninhabitable. Additionally, the data enabled care managers to facilitate transitions from hospitals to appropriate post-acute settings, ensure continuity of prescriptions, and reconnect patients with home health services disrupted during the disaster.
These early applications demonstrate the feasibility of the claims-based identification method in supporting real-time disaster response and informing health system planning for Florida’s Medicaid population with CCCs.
Discussion and Conclusion
This practice report describes Florida Medicaid’s development and application of a claims-based identification method to support preparedness of enrollees with CCCs before emergencies and characterizes the technology-dependent Medicaid population with CCCs. By adapting a validated classification system and operationalizing it within the state’s Medicaid data infrastructure, Florida was able to generate actionable, real-time information to assist health plans and emergency managers in preparing for and responding to natural disasters.
Findings from this report highlight the scale and complexity of the Medicaid population with CCCs, who made up over 7% of all Florida Medicaid enrollees. Notably, nearly 1 in 5 of these individuals are dependent on medical technology, placing them at significant risk during power outages and service disruptions commonly caused by hurricanes.
Geographic analysis revealed that rural and coastal areas, which are often the hardest hit by severe storms, have high concentrations of these medically vulnerable populations. These regions have been repeatedly impacted by major hurricanes in recent years, including Category 5 Hurricane Michael in 2018, Category 3 Hurricane Idalia in 2023, Category 4 Hurricane Helene in 2024, and Hurricane Milton shortly after, causing widespread flooding, infrastructure damage, and compounding recovery challenges. The hurricanes contributed to sustained disruptions in health care access and infrastructure. In 2022, Hurricane Ian struck Central and Southwest Florida as a powerful Category 4 storm, causing catastrophic storm surge, widespread flooding, and prolonged power outages. Florida’s experience demonstrates that Medicaid administrative data can be leveraged not only to support disaster response but also to inform broader health system planning and policy. By identifying where high-risk populations live and the specific services they require, state agencies and health plans can better allocate resources, design targeted interventions, and strengthen care continuity during crises. Additionally, this approach may help address systemic gaps in care for people with complex needs, including those living in disaster-prone communities.
This study adds to the literature and has the potential to influence the delivery and support for continuity of care for Medicaid enrollees during disasters. However, several limitations should be noted. The identification method relied on claims data, which may not capture all medically vulnerable individuals, particularly those with incomplete or delayed claims submissions. The method also did not include direct clinical verification of conditions, which could result in some misclassification. Another limitation was the extrapolation of a case-finding algorithm originally validated in a pediatric CCC population. As a result, the application may not have fully captured the diagnostic nuances, care needs, or service utilization patterns unique to pregnant women, disabled adults, and older adults, potentially leading to under- or overidentification in these groups. Additionally, while health plans reported using the data for outreach and care coordination, this study did not assess individual-level outcomes such as mortality, hospitalization, or health deterioration, which limits our ability to evaluate the direct clinical impact of the intervention on disaster-related health outcomes. Despite these limitations, the approach offers a scalable, reproducible solution that can be refined over time and adapted by other states.
In conclusion, the claims-based CCC identification method in this report was feasible to implement and represents a promising strategy to enhance disaster preparedness and health system resilience. Beyond hurricanes, the method may also inform response planning for other emergencies such as wildfire, flooding, or extreme heat events, where continuity of care for medically vulnerable populations is critical. Other states can build on this work to strengthen their own emergency response capabilities and improve care coordination for medically fragile populations. By doing so, public health agencies and health care organizations may better protect vulnerable communities before, during, and after disasters.
Implications for Policy & Practice
Medicaid programs can leverage existing claims data to proactively identify and map enrollees with complex chronic conditions (CCCs), including those dependent on life-sustaining medical technology.
Providing health plans with real-time, actionable data on medically vulnerable populations is feasible and may enhance their ability to conduct outreach, coordinate care, and maintain service continuity during disasters.
Geographic mapping of CCC prevalence supports targeted resource planning by state and local emergency management agencies, helping prioritize high-risk areas for disaster preparedness efforts.
Other states can adopt this scalable, claims-based identification method to improve health system resilience and emergency response for Medicaid enrollees with high medical needs.
While this case-finding approach was applied to Medicaid data, future policy efforts could expand its utility by incorporating data from Medicare and commercial payers to ensure that emergency preparedness plans address medically complex individuals across all income levels and insurance types.
As these data tools are increasingly adopted by payers and public health agencies, their use must be guided by strong ethical safeguards that prioritize individual autonomy and transparency, especially in Medicaid populations who may be vulnerable to both being overlooked and having their care unduly restricted based on algorithmic classifications.
Footnotes
Jaclyn M. Hall and Madison R. McCraney share the co-first authorship
The authors have no financial disclosures.
The authors have no conflicts of interest.
The authors would like to acknowledge the resilience of Florida’s Medicaid recipients living with complex chronic conditions, whose experiences and challenges inspire ongoing efforts to improve health system responsiveness, care coordination, and emergency preparedness for medically vulnerable populations.
Human Participant Compliance Statement: This project involved the analysis of de-identified administrative claims and enrollment data collected for the purposes of Medicaid program operations and emergency preparedness. As such, it was determined to be public health practice and not human subjects research. Additionally, the project was reviewed by the University of Florida Institutional Review Board (IRB) and determined to be nonhuman exempt from IRB oversight (NH00047073). No identifiable personal health information was accessed or disclosed outside of authorized program operations.
Contributor Information
Jaclyn M. Hall, Email: jaclynha@ufl.edu.
Madison R. McCraney, Email: mm16bm@med.fsu.edu.
Christina A. Vincent, Email: christinavincent.mph@gmail.com.
Peyton A. Lurk, Email: peyton.lurk@gmail.com.
Kristen Erichsen, Email: kristenerichsen@gmail.com.
Choeeta Chakrabarti, Email: cchakrabarti@fsu.edu.
Rahma S. Mkuu, Email: rmkuu@ufl.edu.
Christopher R. Cogle, Email: christopher.cogle@medicine.ufl.edu.
References
- 1.Skinner HG, Coffey R, Jones J, Heslin KC, Moy E. The effects of multiple chronic conditions on hospitalization costs and utilization for ambulatory care sensitive conditions in the United States: a nationally representative cross-sectional study. BMC Health Serv Res. 2016;16(1):77. doi: 10.1186/s12913-016-1304-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tomio J, Sato H, Mizumura H. Interruption of medication among outpatients with chronic conditions after a flood. Prehospital Disaster Med. 2010;25(1):42-50. doi: 10.1017/s1049023x00007652 [DOI] [PubMed] [Google Scholar]
- 3.Gerteis J, Izrael D, Deitz D, LeRoy L, Ricciardi R, Miller T, et al. Multiple Chronic Conditions Chartbook. Agency for Healthcare Research and Quality; 2014:Q14-0038. [Google Scholar]
- 4.Bell SA, Donnelly JP, Li W, Davis MA. Hospitalizations for chronic conditions following hurricanes among older adults: a self-controlled case series analysis. J Am Geriatr Soc. 2022;70(6):1695-1703. doi: 10.1111/jgs.17702 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.McQuade L, Merriman B, Lyford M, Nadler B, Desai S, Miller R, et al. Emergency department and inpatient health care services utilization by the elderly population: hurricane Sandy in the state of New Jersey. Disaster Med Public Health Prep. 2018;12(6):730-738. doi: 10.1017/dmp.2018.1 [DOI] [PubMed] [Google Scholar]
- 6.Sands LP, Do Q, Du P, Pruchno R. Peritraumatic stress from a disaster increases risk for onset of chronic diseases among older adults. Innov Aging. 2022;6(1):igab052. doi: 10.1093/geroni/igab052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Disaster recovery plan. Florida agency for health care administration; 2018.
- 8.Waddell SL, Jayaweera DT, Mirsaeidi M, Beier JC, Kumar N. Perspectives on the health effects of hurricanes: a review and challenges. Int J Environ Res Public Health. 2021;18(5):2756. doi: 10.3390/ijerph18052756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Feudtner C, Feinstein JA, Zhong W, Hall M, Dai D. Pediatric complex chronic conditions classification system version 2: updated for ICD-10 and complex medical technology dependence and transplantation. BMC Pediatr. 2014;14:199. doi: 10.1186/1471-2431-14-199 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Klein RJ, Schoenborn CA. Age adjustment using the 2000 projected U.S. Population: (583772012-001) [Internet]. 2001. https://doi.apa.org/doi/10.1037/e583772012-001. Accessed 2025 May 26.