Prediction of disease severity in patients with febrile illnesses in resource-limited settings
<p>Febrile illnesses present unique challenges for health systems with scarce resources. Large volumes of mostly self-limiting diseases accompanied by high case-fatality rates for the small proportion of serious infections mean that risk stratification tools must have high sensitivities and/or...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2023
|
| Subjects: |
| _version_ | 1826310993953685504 |
|---|---|
| author | Chandna, A |
| author2 | Turner, P |
| author_facet | Turner, P Chandna, A |
| author_sort | Chandna, A |
| collection | OXFORD |
| description | <p>Febrile illnesses present unique challenges for health systems with scarce resources. Large volumes of mostly self-limiting diseases accompanied by high case-fatality rates for the small proportion of serious infections mean that risk stratification tools must have high sensitivities and/or specificities according to their proposed contexts of use. Unfortunately, accuracy and reliability of existing tools are sub-optimal and they are often impractical for deployment in resource-limited settings.</p>
<p>This thesis explores the development and application of prediction tools for the management of febrile illnesses across a range of resource-constrained settings in South and Southeast Asia. It aims to combine the best prediction model science with a pragmatic field-based reality to address locally-important health issues. Recognising that evaluation of prediction tools should be set in their intended contexts of use, each analysis is framed in a particular clinical use-case, yet draws upon approaches to allow exploration of the generalisability of the findings.</p>
<p>Using a variety of research methodologies, this thesis identifies prognostic factors amongst febrile children presenting to different levels of the health system, and uses these data to externally validate existing severity scores and develop new clinical prediction models suitable for resource-limited settings. Prospective work evaluates the relative contributions of clinical and biomarker-based approaches for the referral of children with respiratory infections from a resource-limited community setting on the Thailand-Myanmar border, whilst retrospective work develops a prognostic model for critically ill children on admission to a paediatric intensive care unit in northern Cambodia. Finally, in response to the arrival of the SARS-CoV-2 pandemic, this thesis reports the development and external validation of prognostic models at two sites in India to support the safe outpatient management of patients presenting with moderate Covid-19.</p>
<p>Important challenges and potential solutions to developing prediction tools in resource-limited settings are discussed, including application of the classical prediction paradigm to the assessment of disease severity, comparative analyses of the clinical utility of different models, and the differential importance of various predictors for identifying both patients who are sick at the time of clinical assessment and those whose illnesses will progress later in their disease course.</p> |
| first_indexed | 2024-03-07T08:01:51Z |
| format | Thesis |
| id | oxford-uuid:46cbfae9-91bb-4725-884e-1cf1d4222ad7 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T08:01:51Z |
| publishDate | 2023 |
| record_format | dspace |
| spelling | oxford-uuid:46cbfae9-91bb-4725-884e-1cf1d4222ad72023-10-03T17:00:24ZPrediction of disease severity in patients with febrile illnesses in resource-limited settingsThesishttp://purl.org/coar/resource_type/c_db06uuid:46cbfae9-91bb-4725-884e-1cf1d4222ad7Tropical medicinePublic healthWorld healthMedicineEnglishHyrax Deposit2023Chandna, ATurner, PLubell, YPerera-Salazar, RDay, NTurner, CCarrol, EWertheim, H<p>Febrile illnesses present unique challenges for health systems with scarce resources. Large volumes of mostly self-limiting diseases accompanied by high case-fatality rates for the small proportion of serious infections mean that risk stratification tools must have high sensitivities and/or specificities according to their proposed contexts of use. Unfortunately, accuracy and reliability of existing tools are sub-optimal and they are often impractical for deployment in resource-limited settings.</p> <p>This thesis explores the development and application of prediction tools for the management of febrile illnesses across a range of resource-constrained settings in South and Southeast Asia. It aims to combine the best prediction model science with a pragmatic field-based reality to address locally-important health issues. Recognising that evaluation of prediction tools should be set in their intended contexts of use, each analysis is framed in a particular clinical use-case, yet draws upon approaches to allow exploration of the generalisability of the findings.</p> <p>Using a variety of research methodologies, this thesis identifies prognostic factors amongst febrile children presenting to different levels of the health system, and uses these data to externally validate existing severity scores and develop new clinical prediction models suitable for resource-limited settings. Prospective work evaluates the relative contributions of clinical and biomarker-based approaches for the referral of children with respiratory infections from a resource-limited community setting on the Thailand-Myanmar border, whilst retrospective work develops a prognostic model for critically ill children on admission to a paediatric intensive care unit in northern Cambodia. Finally, in response to the arrival of the SARS-CoV-2 pandemic, this thesis reports the development and external validation of prognostic models at two sites in India to support the safe outpatient management of patients presenting with moderate Covid-19.</p> <p>Important challenges and potential solutions to developing prediction tools in resource-limited settings are discussed, including application of the classical prediction paradigm to the assessment of disease severity, comparative analyses of the clinical utility of different models, and the differential importance of various predictors for identifying both patients who are sick at the time of clinical assessment and those whose illnesses will progress later in their disease course.</p> |
| spellingShingle | Tropical medicine Public health World health Medicine Chandna, A Prediction of disease severity in patients with febrile illnesses in resource-limited settings |
| title | Prediction of disease severity in patients with febrile illnesses in resource-limited settings |
| title_full | Prediction of disease severity in patients with febrile illnesses in resource-limited settings |
| title_fullStr | Prediction of disease severity in patients with febrile illnesses in resource-limited settings |
| title_full_unstemmed | Prediction of disease severity in patients with febrile illnesses in resource-limited settings |
| title_short | Prediction of disease severity in patients with febrile illnesses in resource-limited settings |
| title_sort | prediction of disease severity in patients with febrile illnesses in resource limited settings |
| topic | Tropical medicine Public health World health Medicine |
| work_keys_str_mv | AT chandnaa predictionofdiseaseseverityinpatientswithfebrileillnessesinresourcelimitedsettings |