How are children with medical complexity being identified in epidemiological studies? A systematic review

Patrícia Vicente Oliveira · Carla C.Enes · Luciana B.Nucci

Abstract Background There are different definitions to identify/classify children with medical complexity (CMC).We aimed to investigate and describe the definitions used to classify CMC in epidemiological studies.Methods PubMed,SciELO,LILACS,and EMBASE were searched from 2015 to 2020 (last updated September 15th,2020) for original studies that presented the definition used to classify/identify CMC in the scientific research method.We applied the Preferred Reporting Items for Systematic Reviews and Meta-Analyses methodology.From the included studies,the following were identified: first author,year of publication,design,population,study period,the definition of CMC used,limitations,and strengths.Results Nine hundred and sixty-seven records were identif eid in the searched databases,and 42 met the inclusion criteria.Of the 42 studies included,the four most frequent definitions used in the articles included in this review were classification of CMC into nine diagnostic categories based on the International Classification of Diseases,Ninth Revision (ICD-9) (35.7%,15 articles);update of the previous classification for ICD-10 codes with the inclusion of other conditions in the definition(21.4%,nine articles);definition based on a medical complexity algorithm for classification (16.7%,seven articles);and a risk rating system (7.1%,three articles).Conclusions CMC definitions using diagnostic codes were more frequent.However,several limitations were found in its uses.Our research highlighted the need to improve health information systems to accurately characterize the CMC population and promote the provision of comprehensive care.

Keywords Children with medical complexity · Complex chronic conditions · Epidemiologic studies · Pediatrics ·Systematic review

Introduction

The debate over childhood chronic conditions has been influenced in the last decade by the transition from a focus primarily on large groups of children with special health care needs [1],a broad and inclusive definition of children who have,or are at increased risk of,development involving chronic physical,behavioral,or emotional illnesses that require specific health and services and/or in excess of that required by children in general [2],to smaller groups of children with medical complexity (CMC) [1],or children with chronic complex conditions (CCC).The terms CCC and CMC are traditionally used to describe children who are more fragile from a medical point of view and need more intense health care [3].In the present study,the CMC nomenclature will be adopted.

The concept of CMC covers children who may have a congenital or acquired multisystem disease,a severe neurological condition with significant functional impairment,and/or technological dependence on the activities of daily living [3].Studies focusing on the CMC population have been highlighted in the scientific community since this group needs specific care that disproportionately increases pediatric expenses,in addition to demanding specific policies and programmatic interventions that diverge from broader groups of children with special health needs [1].

There are several tools to identify and classify the CMC population,and medical records or data collected from surveys conducted with parents and healthcare professionals may be used.However,the literature points out that classifying children with medical complexity at a population level is a challenge [4].In this context,this systematic review aimed to investigate and describe the definitions used to classify CMC in epidemiological studies,analyzing the strengths and limitations involved in each revised definition.

Methods

Protocol and registration

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [5,6] and registered on the PROSPERO platform (CRD42020207123).The PRISMA protocol checklist is found in Supplementary Table 1.

Search strategies

We applied the SPIDER (sample,phenomenon of interest,design,evaluation,research type) tool [7] to formulate the following research question: “What are the frequently used definitions of CMC in epidemiological studies?”.We included articles whose samples consisted of children and adolescents.The phenomenon of interest was the definition of CMC,the design comprised epidemiological studies,the evaluation was the use of the definition,and the research type included quantitative and mixed methods.

The bibliographic databases consulted were SciELO (Scientific Electronic Library Online),PubMed,LILACS (Latin American and Caribbean Health Sciences Literature),and EMBASE.The last update was performed on September 15,2020.In the databases,Medical Subject Headings (MeSH)and keywords were combined by the Boolean method using the operators “AND” and “OR”.The search strategies can be found in Supplementary Table 2.It is important to highlight that the terms “complex chronic condition” and “medical complexity” are not indexed in MeSH.Considering the importance of using these nomenclature to obtain valid search results,we chose to use combinations of these terms called “keywords” with terms indexed in MeSH,such as“child”‚ “adolescent”‚ and “prevalence”.Descriptors were used in English,Portuguese,and Spanish.

Eligibility criteria

The inclusion criteria of the studies followed the relevance criteria form for inclusion (Supplementary Table 3).The inclusion criteria adopted were as follows: describing or mentioning in the study method the diagnostic criteria used to classify children and/or adolescents with medical complexity (with a clear indication of the definition used),articles with epidemiological characteristics,published between 2015 and 2020,and in English,Portuguese,and/or Spanish.

Exclusion criteria consisted of non-original articles(literature review studies,short or brief communications,experience reports,case reports,or other publications not characterized as an original article),primarily qualitative studies with approaches outside the objectives of this review,and non-epidemiological studies (therapeutic and pharmacological methods,studies that addressed issues of care only from the perspective of social rights,and/or the dilemmas of family care).

Study selection and data extraction

The PRISMA [5,6] flow diagram was adopted for the selection of articles.The steps involved in the following the diagram were as follows: identification of studies (initial sample and elimination of duplicates),screening (reading of titles,reading of abstract and full text reading),and inclusion (obtaining the final sample: articles that met all relevant criteria for inclusion).For screening and inclusion,the studies were evaluated by two reviewers (OPV and NLB)who independently and blindly verified whether they met the relevant criteria for inclusion.The search results were exported to the Rayyan QCRI [8] platform,which assists in the identification,exclusion of duplicate articles,and blinding between reviewers.

Thus,at each screening step of the PRISMA flow diagram protocol,agreement was verified between reviewers,and in cases where there was a lack of agreement,the decision was made by consensus.In the discussion about studies in which there was no agreement between the reviewers,the relevance criteria form for inclusion (Supplementary Table 3) were applied once more,and this time together,a new decision was made on the inclusion or exclusion of the study.The articles included were organized in an Excel spreadsheet containing the following data: first author,year of publication,design,population,study period,definition of CMC used,and limitations.

Synthesis of the results

Descriptive analysis of the results was used: identification of authorship,methodological study design,description of the population,definition used to classify CMC,study period,and limitations.We also include a quantitative mention of citations of CMC definitions from our reviewed sample and the strengths of the definitions.

Results

From the search,967 publications were identified: Pub-Med (n=406),SciELO (n=66),LILACS (n=164) and EMBASE (n=331),with 42 studies included in this review (Fig.1).Supplementary Table 4 presents the identification and synthesis of the included studies.The definitions by Feudtner et al.[9] and their update by Feudtner et al.[10] together accounted for 57.1% of the sample.The results of our systematic review reported a variety of definitions of CMC applied in epidemiological studies.Twelve distinct definitions were identified,eight applied as a single definition [3,9–15] and four definitions were used in combination with others [3,9,15–19] (Table 1).To our knowledge,this study is the first systematic review that evaluated existing definitions for classifying CMC in epidemiological studies.

Fig.1 Preferred reporting items for systematic reviews and meta-analyses flow diagram.SciELO Scientific Electronic Library Online,LILACS Latin American and Caribbean Health Sciences Literature

In one of the first studies on the theme using diagnostic codes,Feudtner et al.developed a list of ICD-9 codes to classify CMC with a comprehensive definition corresponding to any medical condition that lasts for at least 12 months(except when the patient dies) and which involves distinct organ systems or an organ/system severely enough to require specialized pediatric care and probably some period of hospitalization in a tertiary care center [9].The authors divided the conditions that characterize the CMC public into nine categories: (1) neuromuscular;(2) cardiovascular;(3) respiratory;(4) renal;(5) gastrointestinal;(6) hematologic or immunologic;(7) metabolic;(8) other genetic/congenital defects;and (9) malignancy [9].In 2001,the authors made minor modifications to the initial list published in the 2000 study,including the diagnostic code of epilepsy and expanding the scope of the ICD-9 codes to the subcategories of intellectual disability,degeneration and central nervous system disease,and malignancy [17].In our review,15 studies used the definition by Feudtner et al.[20–34].

In 2004,the Clinical Risk Groups were prepared by the 3M Health Information Systems and the National Association of Children’s Hospitals and Institutions,which consist of a demand-based rating system for risk adjustment,assigning each individual to a single mutually exclusive risk group based on clinical and demographic characteristics to predict the future use of health resources [12].In our review,three studies used this definition [35–37].

In 2011,Cohen and other authors presented a conceptual design with four main domains to characterize CMC:(1) needs: children with specialized and continuous care needs;(2) chronic conditions: having one or more chronic conditions with a determined or unknown diagnosis,and the condition and/or its sequelae must be expected potentially throughout life;(3) functional limitations: having severe limitations that may require technological resources (e.g.,tracheostomy,gastrostomy,wheelchair,etc.);and (4) use of health care: children with a high demand for the use of health services (e.g.,prolonged hospitalization,multiple surgeries,and/or continuous care involving a multidisciplinary team,medication resources,etc.),the intensity of the need for health care may also vary over time [3].In this review,one study used Cohen et al.’s definition as the only criterion[38].Therefore,the authors suggest that future research prospectively examine the ideal method to identify CMC and,consequently,recognize patients with high resource use.A combination of administrative and research data may be needed to increase the sensitivity and specificity of a predictive algorithm compared to a gold standard for clinical evaluation and review of medical records [38].

In 2014,Feudtner et al.presented version 2.0 of their classification system published in 2000 by matching the ICD-9 codes to ICD-10 codes and updating the definition of CMC with the inclusion of codes that specify technological dependence,prematurity and injuries resulting from the neonatal period,and transplantation [10].In our review,nine studies [39–47] used this definition to identify CMC.

Additionally,in 2014,Simon et al.published an article presenting the pediatric medical complexity algorithm(PMCA),a new method based on ICD-9 codes to classify children with chronic diseases according to the level of medical complexity [11].The PMCA considers that the characterization of CMC involves significant chronic conditions in two or more bodily systems among the cardiac,craniofacial,dermatological,endocrine,gastrointestinal,genetic,genitourinary,hematological,immunological,mental health,metabolic,musculoskeletal,neurological,ophthalmological,otological,pulmonary/respiratory,and renal health systems;a chronic condition corresponding to a physical,mental,or developmental condition that may last for at least one year;and a condition that demands health resources above the level of a healthy child,requiring treatment to control the condition,which may be episodically or continuously debilitating [11].If these characteristics do not exist,according to the PMCA,the definition of CMC also involves a progressive condition that is associated with deteriorating health with a shortened life expectancy in adulthood;continuous dependence on technology for at least six months(e.g.,tracheostomy,gastrostomy,mechanical ventilation,kidney dialysis,among other technologies);or progressive or metastatic malignant diseases that affect vital function(e.g.,lymphoma,leukemia,brain tumor),excluding those in remission for more than five years [11].Among the studies included in this review,seven [48–54] used the PMCA as the only criterion to characterize CMC.

Furthermore,in some reviewed papers [55–57],definitions were found to identify CMC originating from specific services,such as Children with Special Health Care Needs in the United States [13],the American Complex Care Service(Boston) [14] and the severe motor and intellectual disabilities definition of the Medical Care Dependent Group used in Japan [15].Four studies chose to use two or more definitions to classify CMC [58–61].

Discussion

In our review,it was possible to note that most of the limitations in identifying CMC in epidemiological studies were related to the particularities of diagnosis and complexity status of the study population,which makes it difficult to identify in a comprehensive manner.Therefore,when the analysis of medical complexity is extended to the population level or involves large samples,three challenges arise: (1)the construction of medical complexity is considered differently among parents,physicians,researchers,and other health professionals;(2) individual-level details about the child are not always available in population-level databases;and (3) children with medical complexity have a heterogeneous set of rare health problems,and yet,normally,one problem does not dominate more in prevalence and impact compared to another [4].However,the full understanding of CMC,in addition to the diagnoses that make up the complexity,becomes increasingly relevant for the planning of care for this population.This is because it has been considered that,as the field of complex pediatric care evolves,the goal is to incorporate measures based on the broader view of health care [62] and on the social determinants of health to define and measure the health of the CMC population [63].Commonly,CMC need home care,social assistance,durable technological health equipment,defense of rights for a long period of time,and multidisciplinary assistance involved in their care [64].

In a study that aimed to present and discuss a line of care for CMC health conditions,the authors considered it of fundamental importance to review the definitions to investigate concepts that would collaborate in a more practical way to describe the characteristics,health demands,and needs of this population.CMC,in addition to being recognized as“chronic” children,are mainly distinguished by degrees of complexity in the use of health services,in the technological support necessary to guarantee vital functions in hospital environments or at home,of having their birth and much of their development marked by living in scenarios,such as wards or intensive care units in pediatric hospitals.Thus,for the authors,definitions are fundamental,as they are associated with practices,guide professional training,and guide public policies [65].

Qualitative research makes it possible to understand the profile of children participating in CMC status,identifying the social determinants of health involved in the characterization of this population.In the care monitoring stage,it is possible to better understand the main needs of this population (technological resources,medication,human resources),analyze the patient's trajectory through different services,verify their health outcomes (percentage of care to their needs,waiting time for care,impacts of waiting time for the worsening of the chronic condition),and measure caregiver burden,among other aspects.However,how best are CMC identified in large populations? Is it also possible to monitor CMC using database as a main strategy? Answering these questions was our main motivation for this study.

Considering the results of our literature review,we understand that many advances in databases are needed to better identify CMC in large samples.We believe that considering the database to support clinical decisions (analyze the data,aiming at health decision-making),increasing the number of variables available,and training health professionals to record the information are the fundamental for the identification of CMC at the population level.In addition,sharing data across multiple sources of information can help to identify the CMC population and their care needs.

Considering the reviewed studies,the variables to be implemented that could help better identify CMC,we can cite the technological assistance that the child uses;the amount and discrimination of daily health procedures;use of multi-professional assistance (e.g.,speech,language and hearing sciences;physiotherapy;occupational therapy;psychology);transfer of the patient to another service;previous comorbidities and those acquired during the course of the disease;history of readmissions;sociodemographic characteristics;neonatal history and neuropsychomotor development (from parents’ and health professionals’ views);and history of services that the child has attended.

On the issue of the relevance of the database for decision-making in health,for complex patients,there is a successful example in the literature: the Medical Information Mart for Intensive Care (MIMIC)-III.It is an open database built through different sources of information that includes information of patients admitted to intensive care units in a large tertiary hospital.It incorporated important variables for care and provided the unfolding of several relevant studies.MIMIC has been an example of the use of artificial intelligence in a hospital context [66,67].

CMC definitions using diagnostic codes were more frequent.However,several limitations were found in its uses,mainly involving errors/inconsistencies in the use of codes,difficulty in accurately measuring the entire CMC population,and lack of more detailed information about the child's complexity status.Our research highlighted the need to improve health information systems to accurately characterize the CMC population and promote the provision of comprehensive care.We believe that improving the classification of the CMC population at the population level can be explored in the health information systems themselves if service providers see the data as a powerful tool to analyze and decision-making in health,overcoming the vision of the associated use only for billing procedures/operational.

Supplementary InformationThe online version contains supplementary material available at https:// doi.org/ 10.1007/ s12519-022-00672-9.

Author contributionsOPV contributed to conceptualization,methodology,formal analysis,investigation,data curation,visualization,and writing of the original draft.ECC contributed to formal analysis,visualization,reviewing,and editing.NLB contributed to conceptualization,methodology,formal analysis,investigation,data curation,visualization,project administration,supervision,reviewing,and editing.All the authors approved the final version of the manuscript.

Data availabilityAll data generated or analyzed during the current study are included in this published article (and its supplementary information files).

Declarations

Ethical approvalNot applicable.

Conflict of interestNo financial or non-financial benefits have been received or will be received from any party related directly or indirectly to the subject of this article.The authors have no conflict of interest to declare.