User feedback on the NHS test & Trace Service during COVID-19: The use of machine learning to analyse free-text data from 37,914 England adults.

Objectives: The UK government's approach to the pandemic relies on a test, trace and isolate strategy, mainly implemented via the digital NHS Test & Trace Service. Feedback on user experience is central to the successful development of public-facing Services. As the situation dynamically changes and data accumulate, interpretation of feedback by humans becomes time-consuming and unreliable. The specific objectives were to 1) evaluate a human-in-the-loop machine learning technique based on structural topic modelling in terms of its Service ability in the analysis of vast volumes of free-text data, 2) generate actionable themes that can be used to increase user satisfaction of the Service. Methods: We evaluated an unsupervised Topic Modelling approach, testing models with 5-40 topics and differing covariates. Two human coders conducted thematic analysis to interpret the topics. We identified a Structural Topic Model with 25 topics and metadata as covariates as the most appropriate for acquiring insights. Results: Results from analysis of feedback by 37,914 users from May 2020 to March 2021 highlighted issues with the Service falling within three major themes: multiple contacts and incompatible contact method and incompatible contact method, confusion around isolation dates and tracing delays, complex and rigid system. Conclusions: Structural Topic Modelling coupled with thematic analysis was found to be an effective technique to rapidly acquire user insights. Topic modelling can be a quick and cost-effective method to provide high quality, actionable insights from free-text feedback to optimize public health Services.

Increasing uptake of National Health Service Health Checks in primary care: a pragmatic randomized controlled trial of enhanced invitation letters in Northamptonshire, England.

BACKGROUND: Uptake of NHS Health Checks (NHSHCs) is sub-optimal. This study aimed to increase their uptake using behaviourally informed invitation letters. METHOD: Patients registered with 6 general practices in Northamptonshire, England who were eligible for an NHSHC between 10 February 2014 and 31 January 2015 were randomized monthly, using a random number generator, to three trial arms: control (standard invitation), sunk costs (resources already allocated) and counterargument (against common barriers to attendance). The outcome measure was uptake of NHSHC by 12 weeks after 31 January. RESULTS: In total, 6331 patients were randomized. After exclusions, due to ineligibility for the NHSHC, data were analysed for N = 6313 patients: N = 2123 control; N = 2085 counterargument; N = 2105 sunk costs. Overall, 2364 (37.45%) patients attended an NHSHC. Both intervention letters increased uptake compared to control, by 5.46% using counterargument (adjusted odds ratio (AOR) 1.32, CI 1.162-1.51, p < 0.001) and 4.33% using sunk costs (AOR 1.246, CI 1.10-1.42, p < 0.001), with no significant difference between the two. CONCLUSION: Behaviourally informed invitation letters, containing sunk costs or counterargument messages, can improve the uptake of NHSHCs. The trial was registered with the International Standard Randomised Controlled Trial Registration Number Scheme (ISRCTN57110614).

Improving child weight management uptake through enhanced National Child Measurement Programme parental feedback letters: A randomised controlled trial.

This single-blind, pragmatic, cluster randomised controlled trial aims to investigate uptake of children's weight management services in response to enhanced National Child Measurement Programme (NCMP) letters providing weight status feedback to parents in three English counties in 2015. Parents of 2642 overweight or very overweight (obese) children aged 10-11 years received an intervention or control letter informing them of their child's weight status. Intervention letters included (i) a visual tool to help weight status recognition, (ii) a social norms statement, and for very overweight children, (iii) a prepopulated booking form for weight management services. The primary outcome was weight management service enrolment. Additional outcome measures included attendance at and contact made with weight management services, and a number of self-report variables. A small effect was observed, with intervention parents being significantly more likely to enrol their children in weight management services (4.33% of Intervention group) than control parents (2.19% of Control group) in both unadjusted (OR = 2.08, p = .008) and adjusted analyses (AOR = 2.48, p = .001). A similar picture emerged for contact with services (4.80% Intervention vs. 2.41% Control; OR = 2.10, p = .003; AOR = 2.46, p < .001) and attendance at services, although group differences in the latter measure were not significant after corrections for multiple comparisons (1.89% Intervention vs. 1.02% Control; AOR = 2.11, p = .047). No effects were found on self-report variables. Theoretically informed weight status feedback letters appear to be an effective strategy to improve enrolment in paediatric weight management services.

Outcomes in HIV-infected adults with tuberculosis at clinics with and without co-located HIV clinics in Botswana.

SETTING: Gaborone, Botswana. OBJECTIVE: To determine if starting anti-tuberculosis treatment at clinics in Gaborone without co-located human immunodeficiency virus (HIV) clinics would delay time to highly active antiretroviral therapy (HAART) initiation and be associated with lower survival compared to starting anti-tuberculosis treatment at clinics with on-site HIV clinics. DESIGN: Retrospective cohort study. Subjects were HAART-naïve, aged ≥ 21 years with pulmonary tuberculosis (TB), HIV and CD4 counts ≤ 250 cells/mm(3) initiating anti-tuberculosis treatment between 2005 and 2010. Survival at completion of anti-tuberculosis treatment or at 6 months post-treatment initiation and time to HAART after anti-tuberculosis treatment initiation were compared by clinic type. RESULTS: Respectively 259 and 80 patients from clinics without and with on-site HIV facilities qualified for the study. Age, sex, CD4, baseline sputum smears and loss to follow-up rate were similar by clinic type. Mortality did not differ between clinics without or with on-site HIV clinics (20/250, 8.0% vs. 8/79, 10.1%, relative risk 0.79, 95%CI 0.36-1.72), nor did median time to HAART initiation (respectively 63 and 66 days, P = 0.53). CONCLUSION: In urban areas where TB and HIV programs are separate, geographic co-location alone without further integration may not reduce mortality or time to HAART initiation among co-infected patients.

Men who have sex with men: estimating the size of at-risk populations in London primary care trusts.

The objective of this study was to estimate local at-risk populations of men who have sex with men (MSM) in London primary care trusts (PCTs) to inform the commissioning of targeted health interventions. Estimated population size and prevalence of diagnosed HIV in MSM in all of London were calculated using data from the British National Survey of Sexual Attitudes and Lifestyles (NATSAL), Greater London Authority population estimates and the annual survey of diagnosed MSM (Survey of Prevalent HIV Infections Diagnosed [SOPHID]). Estimated MSM population sizes at the PCT level were calculated using un-weighted and SOPHID-weighted methods and methods discussed. Four-fifths of MSM with diagnosed HIV infection in Greater London lived in inner London. Estimated population size of MSM 16-44 years in inner London was 66,000; estimated overall prevalence of diagnosed HIV infection among MSM was 9.5%. Our models show substantial variation at the PCT level between the two methods. Using the SOPHID-weighted method MSM account for up to 16% of the male population in some London PCTs, compared with as low as 3% in others. We provide a novel method of estimating at-risk MSM populations living in inner London PCTs indicating that proportions of MSM vary widely between PCTs. Significant proportions of MSM among the resident populations in several PCTs warrant inclusion of MSM health needs in core PCT prevention and service programming. In light of data source limitations further validation studies are needed.

Travel for HIV care in England: a choice or a necessity?

OBJECTIVES: The aims of the study were (1) to measure the distance required to travel, and the distance actually travelled, to HIV services by HIV-infected adults, and (2) to calculate the proportion of patients who travelled beyond local services and identify socio-demographic and clinical predictors of use of non-local services. METHODS: The straight-line distance between a patient's residence and HIV services was determined for HIV-infected patients in England in 2007. 'Local services' were defined as the closest HIV service to a patient's residence and other services within an additional 5 km radius. Multivariable logistic regression was used to identify socio-demographic and clinical predictors of accessing non-local services. RESULTS: In 2007, nearly 57 000 adults with diagnosed HIV infection accessed HIV services in England; 42% lived in the most deprived areas. Overall, 81% of patients lived within 5 km of a service, and 8.7% used their closest HIV service. The median distance to the closest HIV service was 2.5 km [interquartile range (IQR) 1.5-4.2 km] and the median actual distance travelled was 4.8 km (IQR 2.5-9.7 km). A quarter of patients used a 'non-local' service. Patients living in the least deprived areas were twice as likely to use non-local services as those living in the most deprived areas [adjusted odds ratio (AOR) 2.16; 95% confidence interval (CI) 1.98-2.37]. Other predictors for accessing non-local services included living in an urban area (AOR 0.77; 95% CI 0.69-0.85) and being diagnosed more than 12 months (AOR 1.48; 95% CI 1.38-1.59). CONCLUSION: In England, 81% of HIV-infected patients live within 5 km of HIV services and a quarter of HIV-infected adults travel to non-local HIV services. Those living in deprived areas are less likely to travel to non-local services.

Trends over calendar time in antiretroviral treatment success and failure in HIV clinic populations.

OBJECTIVE: Effective antiretroviral therapy (ART) has transformed the care of people with HIV, but it is important to monitor time trends in indicators of treatment success and antic future changes. METHODS: We assessed time trends from 2000 to 2007 in several indicators of treatment success in the UK Collaborative HIV Cohort (CHIC) Study, and using national HIV data from the Health Protection Agency (HPA) we developed a model to project future trends. RESULTS: The proportion of patients on ART with a viral load <50 HIV-1 RNA copies/mL increased from 62% in 2000 to 84% in 2007, and the proportion of all patients with a CD4 count <200 cells/microL decreased from 21% to 10%. During this period, the number of patients who experienced extensive triple class failure (ETCF) rose from 147 (0.9%) to 1771 (3.9%). The number who experienced such ETCF and had a current viral load >50 copies/mL rose fromz 118 (0.7%) to 857 (1.9%). Projections to 2012 suggest sustained high levels of success, with a continued increase in the number of patients who have failed multiple drugs but a relatively stable number of such patients experiencing viral loads >50 copies/mL. Numbers of deaths are projected to remain low. CONCLUSIONS: There have been continued improvements in key indicators of success in patients with HIV from 2000 to 2007. Although the number of patients who have ETCF is projected to rise in the future, the number of such patients with viral loads >50 copies/mL is not projected to increase up to 2012. New drugs may be needed in future to sustain these positive trends.

Immunosuppression among HIV-1-positive patients attending for care: experience from two large HIV centres in the United Kingdom.

OBJECTIVES: The aim of the study was to describe the prevalence of and examine the factors associated with immunosuppression (CD4 < 200 cells/microL) among HIV-infected patients attending two large inner London treatment centres. METHODS: Patients attending for care who had a CD4 count < 200 cells/microL during a 6-month period (1 January to 30 June 2007) were identified from the UK national CD4 surveillance database. Corresponding case notes were reviewed and factors associated with the most recent immunosuppressive episode examined. Patients either previously had a CD4 count > 200 cells/microL at any time under follow-up which had decreased (group A) or never had a CD4 count > 200 cells/microL (group B; late presenters). RESULTS: Of 4589 patients, 10.2% (467) had at least one CD4 count < 200 cells/microL. In group A (60.1% of patients), 70.4% were not receiving antiretroviral therapy (ART) at the time at which the CD4 count fell to < 200 cells/microL. Reasons included: treatment interruption (TI; 32.6%), patient declined ART (20.2%), infrequent attendance (19.1%), physician delay in offer (23.1%) and transient CD4 cell count decrease (3.9%). Among those receiving ART, one in three had poor adherence. In group B, 92.3% had started ART after presentation: most had recently started and were responding virologically. AIDS-defining diagnoses occurred in the year preceding the decrease in CD4 cell count in 12.6% of patients in group A and 33.3% of those in group B. CONCLUSION: The majority of patients became immunosuppressed while under care. Our findings suggest that, in addition to strategies aimed at earlier diagnosis, there are further opportunities to reduce severe immunosuppression in patients already attending for HIV care.

Primary care consultations and costs among HIV-positive individuals in UK primary care 1995-2005: a cohort study.

OBJECTIVES: To investigate the role of primary care in the management of HIV and estimate primary care-associated costs at a time of rising prevalence. METHODS: Retrospective cohort study between 1995 and 2005, using data from general practices contributing data to the UK General Practice Research Database. Patterns of consultation and morbidity and associated consultation costs were analysed among all practice-registered patients for whom HIV-positive status was recorded in the general practice record. RESULTS: 348 practices yielded 5504 person-years (py) of follow-up for known HIV-positive patients, who consult in general practice frequently (4.2 consultations/py by men, 5.2 consultations/py by women, in 2005) for a range of conditions. Consultation rates declined in the late 1990s from 5.0 and 7.3 consultations/py in 1995 in men and women, respectively, converging to rates similar to the wider population. Costs of consultation (general practitioner and nurse, combined) reflect these changes, at pound100.27 for male patients and pound117.08 for female patients in 2005. Approximately one in six medications prescribed in primary care for HIV-positive individuals has the potential for major interaction with antiretroviral medications. CONCLUSION: HIV-positive individuals known in general practice now consult on a similar scale to the wider population. Further research should be undertaken to explore how primary care can best contribute to improving the health outcomes of this group with chronic illness. Their substantial use of primary care suggests there may be potential to develop effective integrated care pathways.

Trends in HIV testing and recording of HIV status in the UK primary care setting: a retrospective cohort study 1995-2005.

OBJECTIVES: To provide nationally representative data on trends in HIV testing in primary care and to estimate the proportion of diagnosed HIV positive individuals known to general practitioners (GPs). METHODS: We undertook a retrospective cohort study between 1995 and 2005 of all general practices contributing data to the UK General Practice Research Database (GPRD), and data on persons accessing HIV care (Survey of Prevalent HIV Infections Diagnosed). We identified all practice-registered patients where an HIV test or HIV positive status is recorded in their general practice records. HIV testing in primary care and prevalence of recorded HIV positive status in primary care were estimated. RESULTS: Despite 11-fold increases in male testing and 19-fold increases in non-pregnant female testing between 1995 and 2005, HIV testing rates remained low in 2005 at 71.3 and 61.2 tests per 100,000 person years for males and females, respectively, peaking at 162.5 and 173.8 per 100,000 person years at 25-34 years of age. Inclusion of antenatal tests yielded a 129-fold increase in women over the 10-year period. In 2005, 50.7% of HIV positive individuals had their diagnosis recorded with a lower proportion in London (41.8%) than outside the capital (60.1%). CONCLUSION: HIV testing rates in primary care remain low. Normalisation of HIV testing and recording in primary care in antenatal testing has not been accompanied by a step change in wider HIV testing practice. Recording of HIV positive status by GPs remains low and GPs may be unaware of HIV-related morbidity or potential drug interactions.

The impact in the UK of the Central and Eastern European HIV epidemics.

Despite increasing migration, the impact of HIV epidemics from Central and Eastern Europe (C&EE) on the UK HIV epidemic remains small. C&EE-born adults comprised 1.2% of adults newly diagnosed with HIV in the UK between 2000 and 2007. Most C&EE-born women probably acquired their infection heterosexually in C&EE. In contrast, 59% of C&EE-born men reported sex with men, half of whom probably acquired their infection in the UK. Previously undiagnosed HIV prevalence in C&EE-born sexual-health-clinic attendees was low (2007, 0.5%) as was overall HIV prevalence in C&EE-born women giving birth in England (2007, <0.1%). The high proportion of men who have sex with men (MSM) suggests under-reporting of this group in C&EE HIV statistics and/or migration of MSM to the UK. In addition to reducing HIV transmission in injecting drug users, preventative efforts aimed at C&EE-born MSM both within their country of origin and the UK are required.

The changing epidemiology of diagnosed prevalent HIV infections in England: greatest impact on the London environs.

Data from the 1997-2004 Surveys of Prevalent HIV Infections Diagnosed were analysed by three geographical areas of residence and treatment to describe the heterogeneous growth of the HIV epidemic in England and provide projections to 2007. Between 1997 and 2004, the number of diagnosed HIV-infected adults resident in England increased by 163% (14,223 to 37,459). Within the 'London environs' the increase was 360% (742 to 3411), within the rest of England 219% (4417 to 14,088) and within London 120% (9064 to 19,960). By 2004, the London environs had the largest proportion of infections acquired through heterosexual sex (and in particular women) and the most recently diagnosed population. Projections indicate over half of diagnosed HIV-infected adults will live outside London by 2007. The epidemiology of diagnosed HIV infection within the London environs is likely to be a predictor of future trends in England overall.

Monitoring the effectiveness of HIV and STI prevention initiatives in England, Wales, and Northern Ireland: where are we now?

Primary and secondary prevention are essential components of the response to HIV and sexually transmitted infections (STIs). We present findings from nationally implemented HIV/STI prevention interventions. In 2003, of those attending STI clinics at least 64% of men who have sex with men (MSM) and 55% of heterosexuals accepted a confidential HIV test; 88% of all HIV infections in women giving birth in England were diagnosed before delivery; 85% of MSM eligible for hepatitis B vaccination received a first dose of vaccine at their first STI clinic attendance; 74% of STI clinic attendees for emergency appointments, and 20% of those for routine appointments were seen within 48 hours of initiating an appointment; the National Chlamydia Screening Programme in England found a positivity of 10% and 13% among young asymptomatic women and men, respectively. Prevention initiatives have seen recent successes in limiting further HIV/STI transmission. However, more work is required if current levels of transmission are to be reduced.

Trends in, and determinants of, HIV testing at genitourinary medicine clinics and general practice in England, 1990-2000.

OBJECTIVES: To describe the trends in and determinants of HIV testing and positivity at genitourinary medicine (GUM) clinics and in general practice (GP) in England between 1990 and 2000. METHODS: Data on all first HIV specimens from GUM and GP clinics and tested at seven sentinel laboratories were related to key demographic, clinical, and behavioural variables. RESULTS: During the observation period, 202 892 eligible first HIV tests were reported. 90% (182 746) of specimens were from GUM clinics, of which 55% were from heterosexuals, 12% from men who have sex with men (MSM), and 3% from injecting drug users (IDU). In contrast, only 3% of GP specimens were from MSM and 13% from IDUs. The total number of first HIV tests increased threefold between 1990 and 2000. Overall, 1.6% of GUM and 0.9% of GP first testers were diagnosed HIV positive. In GUM clinics, HIV positivity was highest among heterosexuals who have lived in Africa (11.7%), MSM (6.9%), and IDUs (2.8%) and lowest among heterosexuals with no other specified risk (0.3%). Consistently lower prevalences were observed in GP settings. HIV positivity among GUM first testers declined in MSM, from 13.6% in 1990 to 5.2% in 2000 (p<0.01), and in IDUs, from 7.5% in 1990 to 2.0% in 2000 (p = 0.03). Prevalence remained constant in the groups heterosexually exposed to HIV infection. CONCLUSIONS: HIV testing in GUM settings increased over the decade, with a concomitant reduction in HIV positivity among MSM and IDUs. Increased testing among heterosexual first testers overall was not associated with declining positivity.