Use this URL to cite or link to this record in EThOS:
Title: Developing a methodology for the evaluation of acute and critical care outcomes in resource-limited settings
Author: Haniffa, Rashan
ISNI:       0000 0004 7430 4920
Awarding Body: University of Oxford
Current Institution: University of Oxford
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Restricted access.
Access from Institution:
The burden of acute and critical illness in LMICs is high, and proportionally higher with poorer outcomes than in HIC. Structured surveillance, enabling systematic evaluation of acute and critical care outcomes, is largely lacking in LMICs. Many tools, including but not limited to prognostic models and decision-support tools, developed in HIC are mostly not validated in LMICs. In addition, acute and critical care skills training, necessary for improving the quality of care and outcomes, is not readily accessible for many healthcare workers. This thesis describes a baseline profile of acute and critical care services in Sri Lanka; the development and implementation of a national, electronic, critical and acute care surveillance system and an assessment of the feasibility of HIC decision-support tools in LMIC settings. It further describes a co-designed, sustainable, national acute and critical care training programme, supported by the surveillance platform. Baseline profile: Overall ICU mortality was 17% but no severity of illness data was available. Overall, only 5.1% of those who had CPR attempted in hospital were alive after 24 hours, with most arrests anticipated by the junior medical team. Only 4.4% of wards use DNAR instructions. The 99 national ICU's had relatively (to other LMICs) good staffing; 790 doctors (1.6 per bed) and 1989 nurses (3.9 per bed, 87.9% ICUs had 1:1 nurse to patient ratio). Evaluation of the applicability of APACHE II was hampered by arterial blood gases and electrolytes being available in only 18.7% and 63.4 % of ICU admissions respectively, and complete case records (for APACHE II) was only available in 1.6% of instances. A surveillance platform for acute and critical care, fusing mobile data entry with visual analytics, was developed and implemented in 56 Sri Lankan hospitals, supporting clinicians in finding ICU beds. The dataset confirmed the low availability of variables commonly used to detect deterioration in acutely unwell ward patients; respiratory rate (65.24 %), mentation (32.89%) and oxygen saturation (23.94%), in a cohort of 16,386 patients. The platform was used for the validation of prognostic models and EWS tools, which showed that the performance of single variable trigger systems was comparable to more complex EWS's regarding identification of at-risk patients. A simpler critical care prognostic model, (TropICS), based on variables more commonly available in LMICs and collected through the platform, was derived and evaluated, and shown to outperform APACHE II in this setting. The platform can also support critical care training; the thesis describes the development, execution, and evaluation of two clinically focused training programmes. A 2-year modular programme in Bangladesh, India and Nepal showed a positive impact on patient outcomes. In Sri Lanka, a peer-delivered, acute and critical care structured training programme was delivered to over 4,500 nurses, physiotherapists and doctors, increasing knowledge and confidence. In summary, the work in this thesis describes a setting-adapted acute and critical care surveillance system, enabling the evaluation of the feasibility and performance of prognostication and decision-support tools, providing a template for LMIC settings. The studies show the importance of evaluation of clinical and benchmarking tools for feasibility and performance, and their adaptation where necessary, prior to their implementation in LMICs. In addition, the studies show that locally developed, sustainable training programmes aimed at improving outcomes in critically ill patients are possible in resource-limited settings. This thesis provides evidence that a clinician-led data platform in a LMIC can provide opportunities to evaluate (and potentially improve) outcomes by an inter-dependent cycle of enhanced information availability, quality improvement, capacity-building, training, and research.
Supervisor: Dondorp, Adrianus ; de Keizer, Nicolette Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available