Acceptability of magnetic resonance imaging for prostate cancer diagnosis with patients and GPs: a qualitative interview study.

BACKGROUND: Magnetic resonance imaging (MRI) of the prostate is a new, more accurate, non-invasive test for prostate cancer diagnosis. AIM: To understand the acceptability of MRI for patients and GPs for prostate cancer diagnosis. DESIGN AND SETTING: Qualitative study of men who had undergone a prostate MRI for possible prostate cancer, and GPs who had referred at least one man for possible prostate cancer in the previous 12 months in West London and Devon. METHOD: Semi-structured interviews, conducted in person or via telephone, were audio-recorded and transcribed verbatim. Deductive thematic analysis was undertaken using Sekhon's Theoretical Framework of Acceptability, retrospectively for patients and prospectively for GPs. RESULTS: Twenty-two men (12 from Devon, age range 47-80 years), two patients' partners, and 10 GPs (6 female, age range 36-55 years) were interviewed. Prostate MRI was broadly acceptable for most patient participants, and they reported that it was not a significant undertaking to complete the scan. GPs were more varied in their views on prostate MRI, with a broad spectrum of knowledge and understanding of prostate MRI. Some GPs expressed concerns about additional clinical responsibility and local availability of MRI if direct access to prostate MRI in primary care were to be introduced. CONCLUSION: Prostate MRI appears to be acceptable to patients. Some differences were found between patients in London and Devon, mainly around burden of testing and opportunity costs. Further exploration of GPs' knowledge and understanding of prostate MRI could inform future initiatives to widen access to diagnostic testing in primary care.

The Clinical Value of Pre-Diagnostic Thrombocytosis for the Detection of Lung Cancer in Primary Care.

Thrombocytosis is a risk marker for lung cancer in primary care. We investigated whether thrombocytosis presents pre-diagnostically for all the histological subtypes of lung cancer and its association with the stage at diagnosis. A matched cohort study used English electronic primary care data linked to the national cancer registry. Patients diagnosed with lung cancer aged ≥40 years with no prior history of malignancy were matched by age, sex, and general practice to five controls without lung cancer. Multivariable logistic regression models quantified the incidence of pre-diagnostic thrombocytosis and advanced-stage diagnoses, adjusting for COPD diagnosis, smoking status, and anti-platelet drug prescriptions. A total of 9504 cases were matched to 45,647 controls, consisting of 3260 (34%) adenocarcinomas (ADC), 2020 (21%) squamous cell carcinomas (SCC), 70 (<1%) large-cell carcinomas (LCC), and 1089 (12%) small-cell lung cancers (SCLC). The patients with lung cancer were 8.9 (95% CI 8.0-9.9) times more likely to exhibit pre-diagnostic thrombocytosis than the controls. The odds ratios were highest for the comparison between SCC and ADC (1.8, 95% CI 1.5-2.1). Thrombocytosis is associated with advanced-stage ADC and SCC but presented equally for early- and advanced-stage SCLC. Pre-diagnostic thrombocytosis may aid in the detection of all the histological subtypes in primary care.

Incidence of 'Low-Risk but Not No-Risk' Features of Cancer Prior to High-Risk Feature Occurrence: An Observational Cohort Study in Primary Care.

Diagnosing cancer may be expedited by decreasing referral risk threshold. Clinical Practice Research Datalink participants (≥40 years) had a positive predictive value (PPV) ≥3% feature for breast, lung, colorectal, oesophagogastric, pancreatic, renal, bladder, prostatic, ovarian, endometrial or laryngeal cancer in 2016. The numbers of participants with features representing a 1-1.99% or 2-2.99% PPV for same cancer in the previous year were reported, alongside the time difference between meeting the ≥3% criteria and the lower threshold criteria. A total of 8616 participants had a PPV ≥3% feature, of whom 365 (4.2%) and 1147 (13.3%), respectively, met 2-2.99% and 1-1.99% criteria in the preceding year. The median time difference was 131 days (Interquartile Range (IQR) 27 to 256) for the 2-2.99% band and 179 days (IQR 58 to 289) for the 1-1.99% band. Results were heterogeneous across cancer sites. For some cancers, participants may progress from presenting lower- to higher-risk features before meeting urgent referral criteria; however, this was not usually the case. The details of specific features across multiple cancer sites will allow for a tailored approach to future reductions in referral thresholds, potentially improving the efficiency of urgent cancer referrals for the benefit both of individuals and the National Health Service (NHS).

Anomalies in the review process and interpretation of the evidence in the NICE guideline for chronic fatigue syndrome and myalgic encephalomyelitis.

Chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME) is a disabling long-term condition of unknown cause. The National Institute for Health and Care Excellence (NICE) published a guideline in 2021 that highlighted the seriousness of the condition, but also recommended that graded exercise therapy (GET) should not be used and cognitive-behavioural therapy should only be used to manage symptoms and reduce distress, not to aid recovery. This U-turn in recommendations from the previous 2007 guideline is controversial.We suggest that the controversy stems from anomalies in both processing and interpretation of the evidence by the NICE committee. The committee: (1) created a new definition of CFS/ME, which 'downgraded' the certainty of trial evidence; (2) omitted data from standard trial end points used to assess efficacy; (3) discounted trial data when assessing treatment harm in favour of lower quality surveys and qualitative studies; (4) minimised the importance of fatigue as an outcome; (5) did not use accepted practices to synthesise trial evidence adequately using GRADE (Grading of Recommendations, Assessment, Development and Evaluations trial evidence); (6) interpreted GET as mandating fixed increments of change when trials defined it as collaborative, negotiated and symptom dependent; (7) deviated from NICE recommendations of rehabilitation for related conditions, such as chronic primary pain and (8) recommended an energy management approach in the absence of supportive research evidence.We conclude that the dissonance between this and the previous guideline was the result of deviating from usual scientific standards of the NICE process. The consequences of this are that patients may be denied helpful treatments and therefore risk persistent ill health and disability.

Effect of bilirubin and Gilbert syndrome on health: cohort analysis of observational, genetic, and Mendelian randomisation associations.

Objectives: To compare associations between the Gilbert syndrome genotype in European populations, measured bilirubin concentrations, genetically predicted bilirubin using this genotype, and a wide range of health outcomes in a large cohort. Design: Cohort study including observational, genetic, and Mendelian randomisation analyses. Setting: 22 centres across England, Scotland, and Wales in UK Biobank (2006-10), with replication in a national Finnish cohort (FinnGen). Participants: 463 060 participants in the UK Biobank were successfully genotyped for a genetic variant (rs887829) that is strongly associated with Gilbert syndrome and 438 056 participants had measured bilirubin concentrations with linked electronic health record data coded using the tenth edition of the International Classification of Diseases. Replication analyses were performed in FinnGen (n=429 209) with linked electronic health record data. Main outcome measures: Odds ratios for the association between serum bilirubin concentrations, rs887829-T homozygosity (the risk genotype for Gilbert syndrome), genetically predicted bilirubin using rs887829-T allele carriage alone, and a wide range of health outcomes recorded in primary and secondary care. Results: 46 189 participants in UK Biobank (about 10%) were homozygous for rs887829-T defining them as having the genotype characterising Gilbert syndrome. However, only 1701 (3%) of this group had a coded diagnosis of Gilbert syndrome. Variation at this locus explained 37.1% of all variation in measured serum bilirubin. In the observational analyses, higher bilirubin concentrations had strong inverse associations with a wide range of outcomes including overall health status, chronic obstructive pulmonary disease, myocardial infarction, and cholesterol measures. These associations were not identified in people with the Gilbert genotype. We identified associations with genetically predicted bilirubin concentrations and biliary and liver pathology (eg, odds ratio for cholelithiasis 1.16 (95% confidence interval 1.12 to 1.20); P=5.7×10-16) and a novel association with pityriasis rosea (1.47 (1.27 to 1.69), P=1.28×10-7). Conclusions: Only 3% of participants who are homozygous for rs887829-T have a recorded diagnosis of Gilbert syndrome. Carriers of this genotype have modest increases in the odds of developing biliary pathology and pityriasis rosea. Evidence from the analyses of genetic data suggests that bilirubin has no likely causal role in protection from cardiovascular disease, chronic obstructive pulmonary disease, or other key healthcare outcomes and therefore represents a poor target for therapeutic intervention for these outcomes.

Can we do better? A qualitative study in the East of England investigating patient experience and acceptability of using the faecal immunochemical test in primary care.

OBJECTIVES: The faecal immunochemical test (FIT) is increasingly used in UK primary care to triage patients presenting with symptoms and at different levels of colorectal cancer risk. Evidence is scarce on patients' views of using FIT in this context. We aimed to explore patients' care experience and acceptability of using FIT in primary care. DESIGN: A qualitative semi-structured interview study. Interviews were conducted via Zoom between April and October 2020. Transcribed recordings were analysed using framework analysis. SETTING: East of England general practices. PARTICIPANTS: Consenting patients (aged ≥40 years) who presented in primary care with possible symptoms of colorectal cancer, and for whom a FIT was requested, were recruited to the FIT-East study. Participants were purposively sampled for this qualitative substudy based on age, gender and FIT result. RESULTS: 44 participants were interviewed with a mean age 61 years, and 25 (57%) being men: 8 (18%) received a positive FIT result. Three themes and seven subthemes were identified. Participants' familiarity with similar tests and perceived risk of cancer influenced test experience and acceptability. All participants were happy to do the FIT themselves and to recommend it to others. Most participants reported that the test was straightforward, although some considered it may be a challenge to others. However, test explanation by healthcare professionals was often limited. Furthermore, while some participants received their results quickly, many did not receive them at all with the common assumption that 'no news is good news'. For those with a negative result and persisting symptoms, there was uncertainty about any next steps. CONCLUSIONS: While FIT is acceptable to patients, elements of communication with patients by the healthcare system show potential for improvement. We suggest possible ways to improve the FIT experience, particularly regarding communication about the test and its results.

Are There Ethnic Differences in Recorded Features among Patients Subsequently Diagnosed with Cancer? An English Longitudinal Data-Linked Study.

We investigated ethnic differences in the presenting features recorded in primary care before cancer diagnosis. METHODS: English population-based cancer-registry-linked primary care data were analysed. We identified the coded features of six cancers (breast, lung, prostate, colorectal, oesophagogastric, and myeloma) in the year pre-diagnosis. Logistic regression models investigated ethnic differences in first-incident cancer features, adjusted for age, sex, smoking status, deprivation, and comorbidity. RESULTS: Of 130,944 patients, 92% were White. In total, 188,487 incident features were recorded in the year pre-diagnosis, with 48% (89,531) as sole features. Compared with White patients, Asian and Black patients with breast, colorectal, and prostate cancer were more likely than White patients to have multiple features; the opposite was seen for the Black and Other ethnic groups with lung or prostate cancer. The proportion with relevant recorded features was broadly similar by ethnicity, with notable cancer-specific exceptions. Asian and Black patients were more likely to have low-risk features (e.g., cough, upper abdominal pain) recorded. Non-White patients were less likely to have alarm features. CONCLUSION: The degree to which these differences reflect disease, patient or healthcare factors is unclear. Further research examining the predictive value of cancer features in ethnic minority groups and their association with cancer outcomes is needed.

Systematic Review and Narrative Synthesis of Economic Evaluations of Prostate Cancer Diagnostic Pathways Incorporating Prebiopsy Magnetic Resonance Imaging.

Context: Prebiopsy magnetic resonance imaging (MRI) of the prostate has been shown to increase the accuracy of the diagnosis of clinically significant prostate cancer. However, evidence is still evolving about how best to integrate prebiopsy MRI into the diagnostic pathway and for which patients, and whether MRI-based pathways are cost effective. Objective: This systematic review aimed to assess the evidence for the cost effectiveness of prebiopsy MRI-based prostate cancer diagnostic pathways. Evidence acquisition: INTERTASC search strategies were adapted and combined with terms for prostate cancer and MRI, and used to search a wide range of databases and registries covering medicine, allied health, clinical trials, and health economics. No limits were set on country, setting, or publication year. Included studies were full economic evaluations of prostate cancer diagnostic pathways with at least one strategy including prebiopsy MRI. Model-based studies were assessed using the Philips framework, and trial-based studies were assessed using the Critical Appraisal Skills Programme checklist. Evidence synthesis: A total of 6593 records were screened after removing duplicates, and eight full-text papers, reporting on seven studies (two model based) were included in this review. Included studies were judged to have a low-to-moderate risk of bias. All studies reported cost-effectiveness analyses based in high-income countries but had significant heterogeneity in diagnostic strategies, patient populations, treatment strategies, and model characteristics. Prebiopsy MRI-based pathways were cost effective compared with pathways relying on ultrasound-guided biopsy in all eight studies. Conclusions: Incorporation of prebiopsy MRI into prostate cancer diagnostic pathways is likely to be more cost effective in than that into pathways relying on prostate-specific antigen and ultrasound-guided biopsy. The optimal prostate cancer diagnostic pathway design and method of integrating prebiopsy MRI are not yet known. Variations between health care systems and diagnostic approaches necessitate further evaluation for a particular country or setting to know how best to apply prebiopsy MRI. Patient summary: In this report, we looked at studies that measured the health care costs and benefits and harms to patients of using prostate magnetic resonance imaging (MRI), to decide whether men need a prostate biopsy for possible prostate cancer. We found that using prostate MRI before biopsy is likely to be less costly for health care services and probably has better outcomes for patients being investigated for prostate cancer. It is still unclear what the best way to use prostate MRI is.

Improving the diagnostic process for patients with possible bladder and kidney cancer: a mixed-methods study to identify potential missed diagnostic opportunities.

BACKGROUND: Patients with bladder and kidney cancer may experience diagnostic delays. AIM: To identify patterns of suboptimal care and contributors of potential missed diagnostic opportunities (MDOs). DESIGN AND SETTING: Prospective, mixed-methods study recruiting participants from nine general practices in Eastern England between June 2018 and October 2019. METHOD: Patients with possible bladder and kidney cancer were identified using eligibility criteria based on National Institute for Health and Care Excellence (NICE) guidelines for suspected cancer. Primary care records were reviewed at recruitment and at 1 year for data on symptoms, tests, referrals, and diagnosis. Referral predictors were examined using logistic regression. Semi-structured interviews were undertaken with 15 patients to explore their experiences of the diagnostic process, and these were analysed thematically. RESULTS: Participants (n = 940) were mostly female (n = 657, 69.9%), with a median age of 71 years (interquartile range 64-77 years). In total, 268 (28.5%) received a referral and 465 (48.5%) had a final diagnosis of urinary tract infection (UTI). There were 33 (3.5%) patients who were diagnosed with cancer, including prostate (n = 17), bladder (n = 7), and upper urothelial tract (n = 1) cancers. Among referred patients, those who had a final diagnosis of UTI had the longest time to referral (median 81.5 days). Only one-third of patients with recurrent UTIs were referred despite meeting NICE referral guidelines. Qualitative findings revealed barriers during the diagnostic process, including inadequate clinical examination, female patients given repeated antibiotics without clinical reviews, and suboptimal communication of test results to patients. CONCLUSION: Older females with UTIs might be at increased risk of MDOs for cancer. Targeting barriers during the initial diagnostic assessment and follow-up might improve quality of diagnosis.

Protocol for a pragmatic cluster randomised controlled trial assessing the clinical effectiveness and cost-effectiveness of Electronic RIsk-assessment for CAncer for patients in general practice (ERICA).

INTRODUCTION: The UK has worse cancer outcomes than most comparable countries, with a large contribution attributed to diagnostic delay. Electronic risk assessment tools (eRATs) have been developed to identify primary care patients with a ≥2% risk of cancer using features recorded in the electronic record. METHODS AND ANALYSIS: This is a pragmatic cluster randomised controlled trial in English primary care. Individual general practices will be randomised in a 1:1 ratio to intervention (provision of eRATs for six common cancer sites) or to usual care. The primary outcome is cancer stage at diagnosis, dichotomised to stage 1 or 2 (early) or stage 3 or 4 (advanced) for these six cancers, assessed from National Cancer Registry data. Secondary outcomes include stage at diagnosis for a further six cancers without eRATs, use of urgent referral cancer pathways, total practice cancer diagnoses, routes to cancer diagnosis and 30-day and 1-year cancer survival. Economic and process evaluations will be performed along with service delivery modelling. The primary analysis explores the proportion of patients with early-stage cancer at diagnosis. The sample size calculation used an OR of 0.8 for a cancer being diagnosed at an advanced stage in the intervention arm compared with the control arm, equating to an absolute reduction of 4.8% as an incidence-weighted figure across the six cancers. This requires 530 practices overall, with the intervention active from April 2022 for 2 years. ETHICS AND DISSEMINATION: The trial has approval from London City and East Research Ethics Committee, reference number 19/LO/0615; protocol version 5.0, 9 May 2022. It is sponsored by the University of Exeter. Dissemination will be by journal publication, conferences, use of appropriate social media and direct sharing with cancer policymakers. TRIAL REGISTRATION NUMBER: ISRCTN22560297.

Workload and workflow implications associated with the use of electronic clinical decision support tools used by health professionals in general practice: a scoping review.

BACKGROUND: Electronic clinical decision support tools (eCDS) are increasingly available to assist General Practitioners (GP) with the diagnosis and management of a range of health conditions. It is unclear whether the use of eCDS tools has an impact on GP workload. This scoping review aimed to identify the available evidence on the use of eCDS tools by health professionals in general practice in relation to their impact on workload and workflow. METHODS: A scoping review was carried out using the Arksey and O'Malley methodological framework. The search strategy was developed iteratively, with three main aspects: general practice/primary care contexts, risk assessment/decision support tools, and workload-related factors. Three databases were searched in 2019, and updated in 2021, covering articles published since 2009: Medline (Ovid), HMIC (Ovid) and Web of Science (TR). Double screening was completed by two reviewers, and data extracted from included articles were analysed. RESULTS: The search resulted in 5,594 references, leading to 95 full articles, referring to 87 studies, after screening. Of these, 36 studies were based in the USA, 21 in the UK and 11 in Australia. A further 18 originated from Canada or Europe, with the remaining studies conducted in New Zealand, South Africa and Malaysia. Studies examined the use of eCDS tools and reported some findings related to their impact on workload, including on consultation duration. Most studies were qualitative and exploratory in nature, reporting health professionals' subjective perceptions of consultation duration as opposed to objectively-measured time spent using tools or consultation durations. Other workload-related findings included impacts on cognitive workload, "workflow" and dialogue with patients, and clinicians' experience of "alert fatigue". CONCLUSIONS: The published literature on the impact of eCDS tools in general practice showed that limited efforts have focused on investigating the impact of such tools on workload and workflow. To gain an understanding of this area, further research, including quantitative measurement of consultation durations, would be useful to inform the future design and implementation of eCDS tools.

Revising the Suspected-Cancer Guidelines: Impacts on Patients' Primary Care Contacts and Costs.

OBJECTIVES: This study aimed to explore the impact of revising suspected-cancer referral guidelines on primary care contacts and costs. METHODS: Participants had incident cancer (colorectal, n = 2000; ovary, n = 763; and pancreas, n = 597) codes in the Clinical Practice Research Datalink or England cancer registry. Difference-in-differences analyses explored guideline impacts on contact days and nonzero costs between the first cancer feature and diagnosis. Participants were controls ("old National Institute for Health and Care Excellence [NICE]") or "new NICE" if their index feature was introduced during guideline revision. Model assumptions were inspected visually and by falsification tests. Sensitivity analyses reclassified participants who subsequently presented with features in the original guidelines as "old NICE." For colorectal cancer, sensitivity analysis (n = 3481) adjusted for multimorbidity burden. RESULTS: Median contact days and costs were, respectively, 4 (interquartile range [IQR] 2-7) and £117.69 (IQR £53.23-£206.65) for colorectal, 5 (IQR 3-9) and £156.92 (IQR £78.46-£272.29) for ovary, and 7 (IQR 4-13) and £230.64 (IQR £120.78-£408.34) for pancreas. Revising ovary guidelines may have decreased contact days (incidence rate ratio [IRR] 0.74; 95% confidence interval 0.55-1.00; P = .05) with unchanged costs, but parallel trends assumptions were violated. Costs decreased by 13% (equivalent to -£28.05, -£50.43 to -£5.67) after colorectal guidance revision but only in sensitivity analyses adjusting for multimorbidity. Contact days and costs remained unchanged after pancreas guidance revision. CONCLUSIONS: The main analyses of symptomatic patients suggested that prediagnosis primary care costs remained unchanged after guidance revision for pancreatic cancer. For colorectal cancer, contact days and costs decreased in analyses adjusting for multimorbidity. Revising ovarian cancer guidelines may have decreased primary care contact days but not costs, suggesting increased resource-use intensity; nevertheless, there is evidence of confounding.

Examining methodology to identify patterns of consulting in primary care for different groups of patients before a diagnosis of cancer: An exemplar applied to oesophagogastric cancer.

BACKGROUND: Current methods for estimating the timeliness of cancer diagnosis are not robust because dates of key defining milestones, for example first presentation, are uncertain. This is exacerbated when patients have other conditions (multimorbidity), particularly those that share symptoms with cancer. Methods independent of this uncertainty are needed for accurate estimates of the timeliness of cancer diagnosis, and to understand how multimorbidity impacts the diagnostic process. METHODS: Participants were diagnosed with oesophagogastric cancer between 2010 and 2019. Controls were matched on year of birth, sex, general practice and multimorbidity burden calculated using the Cambridge Multimorbidity Score. Primary care data (Clinical Practice Research Datalink) was used to explore population-level consultation rates for up to two years before diagnosis across different multimorbidity burdens. Five approaches were compared on the timing of the consultation frequency increase, the inflection point for different multimorbidity burdens, different aggregated time-periods and sample sizes. RESULTS: We included 15,410 participants, of which 13,328 (86.5 %) had a measurable multimorbidity burden. Our new maximum likelihood estimation method found evidence that the inflection point in consultation frequency varied with multimorbidity burden, from 154 days (95 %CI 131.8-176.2) before diagnosis for patients with no multimorbidity, to 126 days (108.5-143.5) for patients with the greatest multimorbidity burden. Inflection points identified using alternative methods were closer to diagnosis for up to three burden groups. Sample size reduction and changing the aggregation period resulted in inflection points closer to diagnosis, with the smallest change for the maximum likelihood method. DISCUSSION: Existing methods to identify changes in consultation rates can introduce substantial bias which depends on sample size and aggregation period. The direct maximum likelihood method was less prone to this bias than other methods and offers a robust, population-level alternative for estimating the timeliness of cancer diagnosis.

Women's priorities towards ovarian cancer testing: a best-worst scaling study.

OBJECTIVE: To investigate the importance of key characteristics relating to diagnostic testing for ovarian cancer and to understand how previous test experience influences priorities. DESIGN: Case 1 best-worst scaling embedded in an online survey. SETTING: Primary care diagnostic testing in England and Wales. PARTICIPANTS: 150 women with ovaries over 40 years old living in England and Wales. METHODS: We used best-worst scaling, a preference-based survey method, to elicit the relative importance of 25 characteristics relating to ovarian cancer testing following a systematic review. Responses were modelled using conditional logit regression. Subgroup analysis investigated variations based on testing history. MAIN OUTCOME MEASURES: Relative importance scores. RESULTS: 'Chance of dying from ovarian cancer' (0.380, 95% CI 0.26 to 0.49) was the most important factor to respondents, closely followed by 'test sensitivity' (0.308, 95% CI 0.21 to 0.40). In contrast, 'time away from usual activities' (-0.244, 95% CI -0.33 to -0.15) and 'gender of healthcare provider' (-0.243, 95% CI -0.35 to -0.14) were least important to respondents overall. Women who had previously undergone testing placed higher importance on certain characteristics including 'openness of healthcare providers' and 'chance of diagnosing another condition' at the expense of reduced emphasis on characteristics such as 'pain and discomfort' and 'time away from usual activities'. CONCLUSIONS: The results clearly demonstrated items at the extreme, which were most and least important to women considering ovarian cancer testing. Differences in priorities by testing history demonstrate an experience effect, whereby preferences adapt over time based on evidence and experience. Acknowledging these differences helps to identify underlying barriers and facilitators for women with no test experience as well as shortcomings of current service based on women with experience.

Machine Learning for Risk Prediction of Oesophago-Gastric Cancer in Primary Care: Comparison with Existing Risk-Assessment Tools.

Oesophago-gastric cancer is difficult to diagnose in the early stages given its typical non-specific initial manifestation. We hypothesise that machine learning can improve upon the diagnostic performance of current primary care risk-assessment tools by using advanced analytical techniques to exploit the wealth of evidence available in the electronic health record. We used a primary care electronic health record dataset derived from the UK General Practice Research Database (7471 cases; 32,877 controls) and developed five probabilistic machine learning classifiers: Support Vector Machine, Random Forest, Logistic Regression, Naïve Bayes, and Extreme Gradient Boosted Decision Trees. Features included basic demographics, symptoms, and lab test results. The Logistic Regression, Support Vector Machine, and Extreme Gradient Boosted Decision Tree models achieved the highest performance in terms of accuracy and AUROC (0.89 accuracy, 0.87 AUROC), outperforming a current UK oesophago-gastric cancer risk-assessment tool (ogRAT). Machine learning also identified more cancer patients than the ogRAT: 11.0% more with little to no effect on false positives, or up to 25.0% more with a slight increase in false positives (for Logistic Regression, results threshold-dependent). Feature contribution estimates and individual prediction explanations indicated clinical relevance. We conclude that machine learning could improve primary care cancer risk-assessment tools, potentially helping clinicians to identify additional cancer cases earlier. This could, in turn, improve survival outcomes.

Real-world effectiveness of pneumococcal vaccination in older adults: Cohort study using the UK Clinical Practice Research Datalink.

BACKGROUND: The 23-valent pneumococcal polysaccharide vaccine (PPV23) is recommended for UK older adults, but how age moderates effectiveness is unclear. METHODS: Three annual cohorts of primary-care patients aged≥65y from the Clinical Practice Research Datalink selected from 2003-5 created a natural experiment (n = 324,804), reflecting the staged introduction of the vaccine. The outcome was symptoms consistent with community-acquired pneumococcal pneumonia (CAP) requiring antibiotics or hospitalisation. We used the prior event rate ratio (PERR) approach to address bias from unmeasured confounders. RESULTS: Vaccinated patients had higher rates of CAP in the year before vaccination than their controls, indicating the potential for confounding bias. After adjustment for confounding using the prior event rate ratio (PERR) method, PPV23 was estimated to be effective against CAP for two years after vaccination in all age sub-groups with hazard ratios (95% confidence intervals) of 0.86 (0.80 to 0.93), 0.74 (0.65 to 0.85) and 0.65 (0.57 to 0.74) in patients aged 65-74, 75-79 and 80+ respectively in the 2005 cohort. Age moderated the effect of vaccination with predicted risk reductions of 8% at 65y and 29% at 80y. CONCLUSIONS: PPV23 is moderately effective at reducing CAP among UK patients aged≥65y, in the two years after vaccination. Vaccine effectiveness is maintained, and may increase, in the oldest age groups in step with increasing susceptibility to CAP.