Identifying eligible patients for Human Immunodeficiency Virus (HIV) testing in UK primary care setting: derivation and internal validation of HIV risk score using a retrospective cohort study and investigation of issues in implementation.

Rumbwere Dube, Benhildah N. (2023). Identifying eligible patients for Human Immunodeficiency Virus (HIV) testing in UK primary care setting: derivation and internal validation of HIV risk score using a retrospective cohort study and investigation of issues in implementation. University of Birmingham. Ph.D.

Preview

RumbwereDube2023PhD_Redacted.pdf
Text - Redacted Version
Available under License All rights reserved.
Download (11MB) | Preview

Abstract

The Acquired Immuno-Deficiency Syndrome (AIDS) and its causative agent, Human Immunodeficiency Virus (HIV) were recognised in the early 1980s. HIV/AIDS is one of the highest contributors to morbidity and a leading cause of mortality, worldwide, making it a public health concern. Diagnosis with HIV during the 1980s and early 1990s was like a death penalty but the life expectancy of the HIV infected individuals has increased over the years and is approaching that for the general population. This is attributed to improvement in management and treatment of people living with HIV/AIDS, the effective use of antiretroviral therapies (ARTs) combined with early HIV diagnosis, because earlier treatment with ARTs is more effective. In addition to enabling individuals to get better health outcomes and optimise their quality of life, early diagnosis of HIV is also important in reducing onward transmission of the disease and reducing healthcare costs. In UK, an action plan was put in plan to eliminate HIV/AIDS by 2025 and this relies on increasing uptake of HIV testing in various settings including primary care, in order to increase early diagnosis.

This thesis developed a prediction model which could be used in primary care to identify patients likely to be HIV positive and to prompt clinicians to offer them HIV testing. It also investigated issues in implementation of a point-of-care alert.

A systematic review was conducted to identify candidate predictor variables to use in the prediction model. Qualitative research was undertaken to find out if GPs used euphemistic terms to record a diagnosis of HIV in primary care records. Using the results from the systematic review and the qualitative research, a cohort study was conducted to derive and internally validate a HIV prediction model. Finally, a systematic review was conducted to elicit clinicians’ views on barriers and facilitators of use of point-of-care alerts to ensure that they are considered in the development of a HIV point-of-care alert.

Demographic, socio-economic, lifestyle/behavioural, and clinical and comorbid conditions that predict risk of HIV infection were identified in this research. The demographic and socio-economic predictors associated with a risk of HIV infection were female gender (HR, 1.19 (CI: 1.13-1.25)), 25-34 years age group (HR, 1.29 (CI: 1.21-1.39)), being of black, (HR 10.95 (CI: 10.08-11.89)) and mixed/other ethnicity, deprivation (HR ranging from 1.3 to 1.85, increasing with deprivation) and living in urban areas (HR, 1.12 (CI: 1-1.25)). Lifestyle predictors were current smoker or ex-smoker (HR, 1.01 (CI:1.01-1.02), drug misuse (HR, 2.25 (CI: 2.01-2.52)) and contact abroad (HR, 2.04 (CI: 1.76-2.36)). Clinical and comorbid conditions included Kaposi’s sarcoma (HR, 171.01 (CI: 89.06-328.37)), pneumocystis carinii (HR, 71.15 (CI: 10.09-501.98)), progressive multifocal leukoencephalopathy (HR, 55.89 (CI: 14.16-220.66)), syphilis (HR, 10.88 (CI: 6.86-17.27)), non-Hodgkin’s lymphoma (HR, 9.31 (CI: 7.04-12.33)), tuberculosis (HR, 2.29 (CI: 1.15-4.55)), cerebral toxoplasmosis abscess (HR, 7.88 (CI: 2.98-20.84)). The other clinical and comorbid predictors were anal cancer or anal intraepithelial dysplasia, aseptic meningitis/encephalitis, oral candidiasis, hepatitis B and C, blood dyscrasia, chronic liver disease, depression and current STI (excluding syphilis) or any previous STI. The C-statistic from the model was 0.74 the optimism adjusted C-slope was 0.990. The sensitivity at 0.25% cut-off was 37% and the specificity was 84%.

The results from the model could develop a risk score to identify patients at high risk of HIV infection in primary care through a point-of-care alert. The identified patients could be offered HIV testing. Barriers and facilitators that affect the use of point-of-care alerts were identified. They fall under the intervention (characteristics of the alert), features pertaining to the setting (GP practice) and person (features related to the clinicians).

This study identified demographic, socio-economic, lifestyle/behavioural, clinical and comorbid conditions that predict risk of HIV infection. These predictors could be used to identify patients at high risk of HIV infection in primary care through a point-of-care alert. The study identified the barriers and facilitators that should be considered to ensure utilisation of a pop-up alert in primary care.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):