Barriers to access and utilization of healthcare by children with neurological impairments and disability in low-and middle-income countries: a systematic review

Background: Neurological impairments (NI) and disability are common among older children in low-and middle-income countries (LMICs). We conducted a systematic review to examine the barriers limiting access and utilization of biomedical and rehabilitative care by children and adolescents with NI in LMICs. Methods: We searched PubMed, Latin America and Caribbean Health Sciences Literature, Global Index Medicus, and Google Scholar for studies published between 01/01/1990 and 14/11/2019 to identify relevant studies. We included all studies reporting on barriers limiting access and utilization of preventive, curative, and rehabilitative care for children aged 0-19 years with NI in five domains: epilepsy, and cognitive, auditory, visual, and motor function impairment. Data from primary studies were synthesized using both qualitative and quantitative approaches. Results: Our literature searches identified 3,258 reports of which 20 were included in the final analysis. Fifteen studies (75.0%) originated from diverse settings in sub-Saharan Africa (SSA). Factors limiting access and utilization of healthcare services in >50% of the studies were: financial constraints (N=17, 85.0%), geographical and physical inaccessibility (N=14, 70.0%), inadequate healthcare resources (N=14, 70.0%), prohibitive culture and beliefs (N=12, 60.0%), and inadequate education/awareness (N=11, 55.0%). Factors reported in <50% of the studies included competing domestic roles (N=4, 20%) and a lack of confidentiality for personal information (N=2, 10.0%). Very few reports were identified from outside Africa preventing a statistical analysis by continent and economic level. Conclusions: Financial constraints, geographic and physical inaccessibility, and inadequate healthcare resources were the most common barriers limiting access and utilization of healthcare services by children with NI in LMICs. PROSPERO registration: CRD42020165296 (28/04/2020)


Introduction
Globally, at least 300 million children live with some form of neurological impairment (NI) or disability, of which >90% originate from low-and middle-income countries (LMICs) (Olusanya et al., 2020). The International Classification of Functioning, Disability and Health, ICF (WHO, 2001), and the International classification of functioning, disability and health: children and youth version: ICF-CY (WHO, 2007), provides a common language and framework to guide research and care of children and youth with NI and disabilities, globally. By integrating both the medical and social models of health and disease (biopsychosocial model (Engel, 1977)), the ICF conceptualizes child functioning as an interaction between the health conditions, environmental factors, and personal factors. Childhood NI and disability not only affects the health and education of the affected individual, but also, directly and indirectly, impacts the family, community, and the society (Patel et al., 2013).
Considering the medical component of the ICF framework, known risk factors of NI and disability in children include adverse pregnancy and birth events such as prematurity and neonatal encephalopathy (Abuga et al., 2021a), infections of the brain including malaria and meningitis, malnutrition, trauma, and poverty (Banks et al., 2017;Scherzer et al., 2012). The affected children, especially those with moderate, severe or multiple disabilities, are vulnerable to infections and accidents which might necessitate hospitalization and subsequent rehabilitation (Moreau et al., 2013). Unfortunately, there is a high treatment gap for NI and disability in LMICs ranging between 50-90% (Patel et al., 2013). The most significant factors limiting the implementation of effective interventions in LMICs include (i) insufficient evidence on the delivery of interventions (ii) inadequate identification of the affected children, and (iii) a shortage of skilled professionals (Patel et al., 2009). Besides, numerous factors limit access or utilization of available preventive, curative, and rehabilitative services by the affected children (WHO, 2011).
Primary preventive services such as nutritional supplementation and immunization have an established role in the prevention of NI (Groce et al., 2014). Vitamin A supplementation and early childhood immunization against measles, rubella, and poliomyelitis substantially reduce the risk of developing NI (Maulik & Darmastadt, 2007), but also reduce morbidity in those that have NI. Secondary prevention involves early screening to identify those already with NI for treatment (or management) to alter the prognosis. Diagnostic tests such as the electroencephalogram (EEG) may help classify seizures and determine treatment (Bassili et al., 2002). Tertiary and quaternary prevention includes treatment and rehabilitation, respectively, to prevent premature mortality, improve functioning, and quality of life. Corrective surgery for children with hearing impairments (Roland et al., 2016) and antiseizure medications for children with epilepsy (Mbuba et al., 2012) are widely documented curative/management options. Rehabilitative services to reduce activity limitation and to improve participation in respective communities include physiotherapy and occupational therapy. However, children from LMICs lack adequate access to the aforementioned continuum of healthcare due to multifarious barriers such as geographical inaccessibility, societal stigma, and financial constraints (WHO, 2011).
A range of barriers, both from the consumer's and provider's perspectives, may hinder uptake of biomedical services by affected children in LMICs ( Figure 1). Poverty in families of children with NI and inadequate government funding limits the prioritization, availability, access, and quality of rehabilitative care (Bright et al., 2018). Besides, prevailing cultural beliefs and societal perceptions shape caregivers' perceptions especially on the aetiology of NI, which affects decisions about the alternatives of care, and ultimately the prognosis of neurodisability (Zuurmond et al., 2019). A 2015 report published by the World Health Organization (WHO) further highlights that people with NI usually experience discrimination at the point of care, which may discourage subsequent seeking of appropriate services (WHO, 2015). Measures such as decentralizing health systems have been proposed to bring services closer to the people and to reduce the geographical distance (distance decay) especially for rural-dwelling populations (Saltman et al., 2007).

Amendments from Version 1
Introduction: We have broadened the introduction section to include the International Classification of Functioning, Disability and Health (ICF) framework to improve the justification of the need for this systematic review. Methods: Following the reviewers' recommendations, we have expanded the search terms to include the terms "OR cerebral palsy" to obtain studies on motor impairments that might have been missed out in the initial searches. We also have explained the checks that were conducted to ensure there was an agreement between the reviewers at the screening stage. In the analysis, more details are provided on the process of identification and classification of barriers limiting access to care by children with neurological impairments and disabilities using data obtained from the primary reports. Results: The total number of studies included in the Results section of the revised paper is 20 compared to 16 reports analysed in the previous version. We also have a higher number of titles/abstracts (N=3,258) compared to the number (N=3,074) reported in the former draft. We also present the quality of the studies report in this section. We have revised Figure 3 to reflect the effect of the expanded searches in the study selection process. We moved Table 1 from the Methods to  Table 3 in the Results, and thereafter, renamed the other tables appropriately both in the text and the captions. We have revised Table 1, Table 2 and Table 4 to include information obtained from the additional reports identified by the expanded searches. As suggested by the reviewers, Figure 4 was removed from the paper in the revised manuscript because it was unnecessary. Discussion: We have revised the discussion section to incorporate interpretation of all studies identified and included in this systematic review. We have replaced the terms "anti-epileptic drugs" with "antiseizure medications" throughout the paper.

REVISED
Some studies from LMICs have separately identified contextual factors that limit access and utilization of existing biomedical services by children with NI. The available evidence on barriers preventing the use of preventive, curative, and rehabilitative services by children with NI in LMICs remains fragmented. We, therefore, conducted a systematic review to collate and classify barriers limiting access and utilization of biomedical services by children and adolescents with NI in LMICs. Synthesized evidence from LMICs is required to inform policy and public health action to ensure equity in access and utilization of healthcare as enshrined in the agenda of the United Nation's sustainable development goals (Pettigrew et al., 2015;Tangcharoensathien et al., 2015).

Reporting guidelines
We used the National Health Service Centre for Review and Dissemination (CRD) recommendations (Booth et al., 2010) and the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines (Abuga et al., 2021b; Moher et al., 2009) to conduct this systematic review. We registered a protocol with the International Prospective Register for Systematic Reviews (PROSPERO), registration ID CRD42020165296 (28 th April 2020).

Search strategy and inclusion criteria
We searched PubMed, Latin America and Caribbean Health Sciences Literature (LILACs), Global Index Medicus, and Google Scholar databases for reports published between 01/01/1990 and 14/11/2019 to identify relevant reports. These searches covered a period through which the burden of NI and disability has significantly increased, globally (Global Research on Developmental Disabilities Collaborators, 2018). The key search terms were 'neurological impairment' and 'access' or 'utilization' and 'healthcare services' as shown in Figure 2, with limits to human studies only.
Eligible studies included those of: (i) children and adolescents aged 0-19 years; (ii) children or adolescents with NI in five domains namely, epilepsy, and impairments in cognitive, hearing, visual, and motor functions; (iii) assessing access and utilization of healthcare services including preventive, curative, and rehabilitative care; and (iv) LMICs as defined by the World Bank (World-Bank, 2019). We excluded systematic reviews, reports, studies on adults, studies conducted in HICs, commentaries, and studies published in languages other than English.
Defining neurological impairment and disability Neurological impairment was defined as a deficit of the central nervous system (CNS) resulting in functional limitation in five domains namely: epilepsy, and cognitive, hearing, visual, and motor impairments (Mung'ala-Odera et al., 2006). Epilepsy was defined according to the International League Against Epilepsy's (ILAE) guidelines as the presence of two or more unprovoked seizures occurring more than 24 hours apart within the previous 12 months (Fisher et al., 2014;Thurman et al., 2011). A child with moderate or severe cognitive impairment refers to a child with a z-score below-2SD or -3SD, respectively based on neuropsychological scores standardized to the normal population. Moderate motor impairment is defined as a difficulty in holding objects, dressing and sitting upright, or ambulant only with help, while severely impaired children include those unable to walk or have no functional use of the hands (WHO, 2011). Moderate hearing impairment refers to a 41-70dB loss in the best ear or difficulty in hearing with a hearing aid; severe impairment refers to greater than 70dB hearing loss or complete loss of hearing in the best ear (WHO, 2020). A child with moderate vision impairment has a visual acuity poorer than 6/18 while those with visual acuity poorer than 6/60 meters are classified as having severe vision impairment (WHO, 2019). However, there was a variation in the actual definitions used in the individual studies and most reports lacked information on NI severity. We, therefore, used these definitions as a formal guideline to verify the actual definitions used in the included reports.
Study selection, data extraction, and quality appraisal Study selection was done in two phases. In the first phase, two reviewers (LM and JA) independently screened the reports identified by the searches by title and abstract for eligibility. The two reviewers compared the lists of the identified reports as to whether they met the predefined inclusion criteria and disagreements were resolved through consensus. In the second phase, both reviewers (LM and JA) examined the fulltext of articles obtained from the first phase against the inclusion and exclusion criteria. Disagreements were resolved through consensus in discussions involving three reviewers (LM, JA and EC). We extracted data relevant for analysis using a pretested data extraction tool designed by the reviewers using guidelines from the PRISMA checklist. Extracted data included author details, study setting, study population, participant characteristics, the type of healthcare services sought, and barriers hindering access or utilization of the services by children with NI or disability. We assessed the quality of each study using the Joanna Briggs Institute (JBI) critical appraisal tools, which are distinct for cross-sectional, cohort, and qualitative studies (Munn et al., 2014).

Synthesis of included reports
Eligible reports identified by the searches and the selection processes were both qualitative and quantitative in design. We, therefore, used a mixed methods review process to synthesize evidence from the eligible studies. The qualitative component of the analytic phase involved the identification and classification of the barriers limiting access to healthcare as described in the primary studies. We grouped all eligible studies based on the domain of NI investigated, and then classified the barriers identified in the primary studies in a three-step process. In step I, three reviewers (LM, EC and JA) independently identified the emerging categories in all the data retrieved from the primary articles. Single or multiple quotes in the primary articles were sufficient because the articles themselves did not similarly present these barriers. In stage II, the three reviewers held a joint meeting to present and discuss all the identified categories, each at a time. Some categories were merged, and others were created to include reviewers' suggestions logically and comprehensively. In stage III, the reviewers held the final joint consensus meeting to review the previously created and discussed categories before writing the final report. This three-step process ensured that the classification process for the barriers identified in the primary articles was valid and reliable. We also identified verbatim excerpts from the eligible primary qualitative studies to represent the voices of the caregivers about their perceptions of the barriers to care for their disabled children. The quantitative component of the analytic phase primarily entailed determining the frequencies and proportions (percentages) of the identified barriers depending on the continent and economic level of the country of origin of the included primary studies. Services sought by children with NI were classified as preventive, curative, or rehabilitative.
Healthcare for children with neurological impairment Rehabilitative services were sought in nine (45.0%) studies, treatment/management reported in seven (35.0%) studies, and preventive care reported in four (20.0%) studies (Table 2) Barriers for healthcare (quantitative analysis) The barriers identified from all eligible studies were classified, analysed, and reported based on seven all-inclusive themes namely: financial constraints, geographical inaccessibility and physical barriers, the inadequacy of healthcare resources, inadequate education/awareness, prohibitive culture/beliefs, competing domestic roles, and a lack of confidentiality/anonymity (Table 3).
The main factors limiting accessing/utilization of healthcare services in >50% of the studies were financial constraints (N=17, 85.0%), geographical inaccessibility (N=14, 70.0%), inadequacy of healthcare resources (N=14, 70.0%), prohibitive culture/beliefs (N=12, 60.0%), and inadequate education/awareness (N=11, 55.0%). Factors reported in less than half of the studies were competing domestic roles (N=4, 20.0%), and the lack of confidentiality/anonymity (N=2, 10.0%). We identified very few reports originating from outside Africa which prevented statistical comparisons of individual reports by continents and economic level of the countries contributing primary studies in the analysis (Table 4).  Geographical inaccessibility described as the proximity of the healthcare facilities from the patient's/caregiver residence. Inaccessibility also included the inability to access healthcare services due to physical disabilities and/or unfavourable infrastructural design of healthcare facilities or transport systems.

Inadequate healthcare resources and quality of care
The attitude of healthcare workers and availability of appropriate services/equipment such as expert consultation, clinical assessment, supply of critical drugs, laboratory equipment, and testing in the healthcare facilities.
6. Confidentiality/anonymity Comprising privacy and protection of patient information of those seeking healthcare.

Discussion
Overall, the main factors hindering access to healthcare by children with NI were financial constraints, geographical and physical inaccessibility, inadequate healthcare resources, prohibitive culture and beliefs, and lack of education or awareness, respectively. Other important but less frequently reported factors were competing for domestic roles for caregivers, and a lack of confidentiality for personal information. Our searches identified more studies in epilepsy, and we obtained fewer reports for the other domains of NI such as hearing impairments. The main form of care reported for epilepsy patients was provision of antiseizure medications; however, there was a lack of newer antiseizure medication and other treatment options such as epilepsy surgery. Rehabilitative services were the most common form of care for those with motor impairments/cerebral palsy and intellectual disability. Most studies originated from Africa and fewer studies were identified from other LMICs complicating valid comparison by continent and level of economic development.
Children with NI from impoverished families could not afford out-of-pocket payments for healthcare, a problem compounded by a lack of health insurance. Many families were forced to neglect healthcare to meet more pressing basic needs such as food and shelter. These can be addressed by expanding the scope of health insurance coverage to reduce out-of-pocket payments for healthcare in LMICs. Children with epilepsy could not regularly access antiseizure medication, a common challenge in developing countries that can be solved by establishing community-based services, outreach programs and reducing the prices of antiseizure medication in settings where they are costly. Indirect costs such as loss of work time have a significant economic impact on these families as time intended to be spent earning an income is used to care for their children. A cycle of poverty and disability might explain the inability of most caregivers to afford expensive primary care for their children (Banks et al., 2017). Wilmshurst and colleagues (Wilmshurst et al., 2014), affirm the finding that management of children with NI is expensive, and the cost of healthcare is unaffordable for many impoverished families. Additional expenses such as public transport and indirect costs such as loss of income were observed in previous studies (Eide et al.,  2015).
Inadequate healthcare resources especially for rehabilitation were common in most African studies. Few rehabilitation experts were working in poorly resourced rehabilitation centres predominantly located in urban centres. The lack of specialist services for visually impaired children reflects a previously reported shortage of 3.7 ophthalmologists per one million people in LMICs, a figure which is substantially low compared to 76. As seen in a previous study (Eide et al., 2015) poor terrain and long geographical distance significantly reduced the likelihood that children with NI from remote areas were able to access healthcare. Also, a lack of environmental modification to cater to those with physical impairment limited access to healthcare in some studies. For instance, there was a shortage of wheelchairs, and public transport systems and hospital facilities lacked provisions for people with physical disabilities. Distance decay, a phenomenon where service utilization reduces with increasing geographical distances from the healthcare facilities can be addressed through decentralizing healthcare and equipping rural health facilities with appropriate healthcare and rehabilitation resources. Additionally, community-based rehabilitation has been recommended to complement the care provided by the existing healthcare systems (Iemmi et al., 2015). While the convention for the rights of people with disabilities, based on the ICF framework, advocates for modification of the environment for the welfare of those with disabilities (WHO, 2001; WHO, 2011), these recommendations have not been implemented in most LMICs. Governments must ensure disability mainstreaming in existing and future infrastructural development (such as the construction of ramps) and partnership with non-governmental/private entities in providing other forms of support (e.g. wheelchairs) for children with disabilities.
In terms of culture and beliefs, our results are similar to those from a Turkish study (Diken, 2006), where mothers who perceived child disability to be a result of curses were more likely to seek traditional interventions. Misconceptions and animistic beliefs on the cause of NI were strongly associated with visiting THP. There is a need to integrate THP into formal healthcare alongside strengthening community-based rehabilitation (Krah et al., 2018). Attitudes from the patient/caregiver, healthcare provider, or societal perspectives were associated with decisions made by primary caregivers or children with NI regarding healthcare. For example, a lack of altruism and discrimination by healthcare providers, and a lack of privacy was observed in sexual and reproductive care for adolescents with disability (

Strengths and limitations
There were few studies identified by our searches, with most reports from Africa and none from South America. Logistical constraints e.g. subscription requirements limited the number of databases we could search as well as the translation of reports identified in other languages than English. Also, we acknowledge that our search strategy might have missed some studies because a range of commonly used etiological and disability terms may have changed over time in the existing literature. This will affect the generalizability of our findings, and specific studies are needed from unrepresented settings. Both quantitative and qualitative studies were eligible for analysis, but study design variability prevented the utility of purely qualitative or quantitative methods in this review. While children and adolescents represent a broad range of age groups possibly with different factors influencing the utility of healthcare, there was no standard reporting of barriers in the primary studies reviewed, which should be standardized in future working groups by expert panels or task forces. However, to the best of our knowledge, this is the first systematic review to synthesize evidence of barriers limiting access and utilization of preventive, curative, and rehabilitative care by children and adolescents with NI in multiple domains in LMICs.

Conclusion
Financial constraints, geographic inaccessibility, inadequate healthcare resources, poor communication/awareness, and cultural barriers were the most ubiquitous barriers limiting access and utilization of healthcare services by children with NI. There were more studies from sub-Saharan Africa where the use of rehabilitative care was more common. Expanding health insurance coverage, improving infrastructure with the decentralization of healthcare, and adequate training and staffing of care facilities, combined with investment in structured health promotion are fundamental steps towards addressing these challenges.

Data availability
All data underlying the results are available as part of the article and no additional source data are required. This important systematic review has been improved by the authors, and they have responded appropriately to the comments I made on the original version. I still would have liked to know what use was made of the quality rating (e.g. whether less weight was given to findings from papers judged to be of low quality). However, it is not necessary for this to be addressed in a further revision. This paper will, I hope, be helpful to service commissioners and providers in considering how to overcome barriers for families in accessing and utilising services for children with neurological impairments and disability.

Is the statistical analysis and its interpretation appropriate? Partly
Are the conclusions drawn adequately supported by the results presented in the review? Partly Thank you very much for reviewing this paper and inviting us to address your comments and suggestions. We have addressed the comments and made a point-by-point response to each one of them as below.

Comment #1
In reviewing this paper, I take note of comments and suggestions by the first reviewer raising valid issues which I agree need to be addressed by the authors.
Reply: Thank you for noting the valid comments and suggestions given by the first reviewer. We have addressed your comments and suggestions, including those given by the first reviewer, and revised the manuscript appropriately.

Comment #2
The systematic review analyses and summarizes different types of studies with different levels of evidence. There needs to be a clearer articulation of the mixed methods review process that has likely been used in conducting the review. This should include an explicit description of the data extraction and data synthesis processes used for the qualitative and quantitative types of studies.

Reply:
We have provided additional information for quantitative and qualitative synthesis approaches on pages 7-8 of the revised paper. We've also described the list of variables that were extracted and the methods which were used in the data synthesis for qualitative and quantitative studies included in this review.

Comment #3
The authors indicate that they assessed the quality of the included studies, but they have not provided a summary of the assessment.
authors' statements concerning how frequently different types of conditions or disorders would be represented in the literature, and such references should be rewritten (e.g. the first sentence of the Discussion, and the second sentence of the Conclusion). Of limitations, first, the authors considered only papers published in English, in only four databases, and did not include grey literature or searches by key author surnames. Second, the search terms for neurodisability were limited mostly to 'impairment', i.e. apart from epilepsy. In the span of time searched (1990-2019), before the publication of the ICF in 2001, a range of etiological and disability terms would have been used. For example, a search on 'cerebral palsy' (not just 'motor impairment') would have yielded a number of additional papers examining barriers to service use in LMICs. It is not clear that the authors did a preliminary check that key expected papers were found by their search strategy, and the resulting focus on epilepsy studies in their results is not surprising. Third, two reviewers independently screened papers on title and abstract for eligibility, but there is no statement as to whether there was any check on agreement at this crucial stage. Fourth, the authors state they examined the quality of studies but how this information was used is not reported. Fifth, there is inadequate description of how the seven themes to classify barriers were derived. Was this drafted in advance, or from the data extracted? Was there any check on the validity/reliability of the classification? What type of evidence was used -were single quotes from a paper sufficient, or was the paper itself required to have presented categories? Thus not enough information is presented so that the systematic review could be repeated by others. Obviously, the authors cannot redo the review at this stage, but a fuller acknowledgement of limitations in the Discussion is necessary.
Overall, the review takes a rather narrow focus, and some of the writing seems to reflect a medical mind-set rather than considering the wider needs of the child in the family and community. For example, the first sentence of the Abstract describes a focus on "survivors of childhood mortality" rather than reflecting the multifactorial background of child neurodisability. Likewise, the first paragraph of the Introduction emphasises hospitalisation from accidents or infection and only moves to prevention and rehabilitation in the second paragraph. The Introduction does not present a very thorough justification of the need for the review, simply saying the evidence concerning barriers is fragmented. To my mind a broader introduction would have been beneficial -for example, I was surprised not to see any reference to the influential overview article by Patel et al. (2013 Arch Dis Child) 1 which also discussed barriers such as lack of evidence-based treatments and lack of identification of disorders.

Comment #7:
Was there any check on the validity/reliability of the classification?
Reply: As highlighted in our response to comment # 6 above, we addressed the inter-observer validity/reliability of our classification approach in a three-step process. We think that there is room for improving this thematic framework in future reviews.

Comment #8:
What type of evidence was used -were single quotes from a paper sufficient, or was the paper itself required to have presented categories? Thus, not enough information is presented so that the systematic review could be repeated by other by others. Obviously, the authors cannot redo the review at this stage, but a fuller acknowledgement of limitations in the Discussion is necessary.

Reply:
As this was a systematic review to provide published data on the subject, single or multiple quotes identified in the primary article were sufficient; the report itself did not have to present the categories as defined by the reviewers as long as the barriers could be inferred from what was already provided. We provide this information in the Methods section (pages 7-8) of the revised manuscript. We also acknowledge the limitations of this approach in the Discussion.

Comment #9:
Overall, the review takes a rather narrow focus, and some of the writing seems to reflect a medical mindset rather than considering the wider needs of the child in the family and community. For example, the first sentence of the Abstract describes a focus on "survivors of childhood mortality" rather than reflecting the multifactorial background of child neurodisability. Likewise, the first paragraph of the Introduction emphasises hospitalisation from accidents or infection and only moves to prevention and rehabilitation in the second paragraph.
Reply: Thank you for raising this important comment. We acknowledge the multidimensional perspectives of disability and how these have evolved. We have revised the abstract and the introduction sections, following your recommendations, to include the multifactorial background to child neurodisability based on the ICF framework, without deviating from the focus of our review, which was to identify barriers limiting access and utilization of biomedical services, including preventive, curative and rehabilitative care.

Comment #10:
The Introduction does not present a very thorough justification of the need for the review, simply saying the evidence concerning barriers is fragmented. To my mind a broader introduction would have been beneficial -for example, I was surprised not to see any reference to the influential overview article by Patel et al. (2013 Arch Dis Child) 1 which also discussed barriers such as lack of evidence-based treatments and lack of identification of disorders.
Reply: Thank you for this suggestion. We have broadened the introduction to justify the need for this systematic review. We have also included ideas from the landmark overview article by Patel