BACKGROUND: Up to 50% of those attending for low-dose computed tomography screening for lung cancer continue to smoke and co-delivery of smoking cessation services alongside screening may maximise clinical benefit. Here we present data from an opt-out co-located smoking cessation service delivered alongside the Yorkshire Lung Screening Trial (YLST). METHODS: Eligible YLST participants were offered an immediate consultation with a smoking cessation practitioner (SCP) at their screening visit with ongoing smoking cessation support over subsequent weeks. RESULTS: Of 2150 eligible participants, 1905 (89%) accepted the offer of an SCP consultation during their initial visit, with 1609 (75%) receiving ongoing smoking cessation support over subsequent weeks. Uptake of ongoing support was not associated with age, ethnicity, deprivation or educational level in multivariable analyses, although men were less likely to engage (adjusted OR (ORadj) 0.71, 95% CI 0.56-0.89). Uptake was higher in those with higher nicotine dependency, motivation to stop smoking and self-efficacy for quitting. Overall, 323 participants self-reported quitting at 4â weeks (15.0% of the eligible population); 266 were validated by exhaled carbon monoxide (12.4%). Multivariable analyses of eligible smokers suggested 4-week quitting was more likely in men (ORadj 1.43, 95% CI 1.11-1.84), those with higher motivation to quit and previous quit attempts, while those with a stronger smoking habit in terms of cigarettes per day were less likely to quit. CONCLUSIONS: There was high uptake for co-located opt-out smoking cessation support across a wide range of participant demographics. Protected funding for integrated smoking cessation services should be considered to maximise programme equity and benefit.
Smoking Cessation , Tobacco Use Disorder , Male , Humans , Smoking Cessation/methods , Community Health Services , Lung , Tomography
BACKGROUND: Cancer incidence increases with age, so some clinical guidelines include patient age as one of the criteria used to decide whether a patient should be referred through the urgent suspected cancer (USC) pathway. Little is known about how strictly GPs adhere to these age criteria and what factors might influence their referral decisions for younger patients. AIM: To understand GPs' clinical decision making for younger patients with concerning symptoms who do not meet the age criteria for USC referral. DESIGN AND SETTING: Qualitative study using in-depth, semi-structured interviews with GPs working in surgeries across England. METHOD: Participants (n = 23) were asked to recall consultations with younger patients with cancer symptoms, describe factors influencing their clinical decisions, and discuss their overall attitude to age thresholds in cancer referral guidelines. A thematic analysis guided by the Framework approach was used to identify recurring themes. RESULTS: GPs' decision making regarding younger patients was influenced by several factors, including personal experiences, patients' views and behaviour, level of clinical concern, and ability to bypass system constraints. GPs weighted potential benefits and harms of a referral outside guidelines both on the patient and the health system. If clinical concern was high, GPs used their knowledge of local systems to ensure patients were investigated promptly even when not meeting the age criteria. CONCLUSION: While most GPs interpret age criteria flexibly and follow their own judgement and experience when making clinical decisions regarding younger patients, system constraints may be a barrier to timely investigation.
BACKGROUND: Parkinson's disease is a multisystem condition that usually presents as a movement disorder in clinical practice. There is no objective method for its diagnosis and therefore the current diagnostic process is based on characteristic clinical signs and symptoms. As the presenting symptoms can be vague and non-specific, there is often a delay in diagnosis leading to mismanagement and delayed treatment initiation. In the UK, GPs identify and initially assess individuals with Parkinson's disease and refer them to specialists for formal diagnosis and treatment initiation. AIM: To use a scoping review to examine the available evidence on the accuracy of Parkinson's disease diagnosis in primary care and to assess the potential for GPs to make a diagnosis and initiate treatment, and hence avoid harmful delays. DESIGN AND SETTING: The scoping methodology as proposed by Westphaln and colleagues that is a modified version of Arksey and O'Malley's original framework was followed. All findings were reported according to PRISMA guidelines for scoping reviews. METHOD: Four databases (EMBASE, PubMed Central, Cochrane, and CINAHL) and references lists of relevant published literature were systematically searched for all types of literature available in English on the accuracy of Parkinsonism or Parkinson's disease diagnosis in primary care. There were no search restrictions placed on countries, type of studies, or age. Two reviewers independently screened titles and abstracts followed by full-text screening. RESULTS: Out of 1844 studies identified, only six studies met the inclusion criteria. Five were from high-income and one from a middle-income nation. Of these, three studies identified significant knowledge gaps of GPs in diagnosing Parkinson's disease using a questionnaire-based assessment. Delay in appropriate referral because of delayed symptom identification was reported in one study. Only one study compared the accuracy of primary care Parkinson's disease diagnosis with that of specialists, and reported that, although specialists' diagnosis showed more sensitivity, GPs had higher specificity in diagnosing Parkinson's disease. However, this study was found to have methodological issues leading to bias in the findings. CONCLUSION: This scoping review shows that there are no well-conducted studies assessing the accuracy of Parkinson's disease diagnoses when made by GPs. This calls for more focused research in this area as diagnostic delays and errors may lead to potentially harmful but preventable delays in treatment initiation resulting in decreased quality of life for individuals with Parkinson's disease.
Parkinson Disease , Humans , Parkinson Disease/diagnosis , Quality of Life , Primary Health Care
OBJECTIVES: In the UK, the number of patients urgently referred for suspected cancer is increasing, and providers are struggling to cope with demand. We explore the potential cost-effectiveness of a new risk prediction test - the PinPoint test - to triage and prioritize patients urgently referred with suspected urological cancers. METHODS: Two simulation models were developed to reflect the diagnostic pathways for patients with (i) suspected prostate cancer, and (ii) bladder or kidney cancer, comparing the PinPoint test to current practice. An early economic analysis was conducted from a UK National Health Service (NHS) perspective. The primary outcomes were the percentage of individuals seen within 2 weeks and health care costs. An exploratory analysis was conducted to understand the potential impact of the Pinpoint test on quality-adjusted life years gained. RESULTS: Across both models and applications, the PinPoint test led to more individuals with urological cancer being seen within 2 weeks. Using PinPoint only to prioritize patients led to increased costs overall, whereas using PinPoint to both triage and prioritize patients led to cost savings. The estimated impact on life years gained/lost was very small and highly uncertain. CONCLUSIONS: Using the PinPoint test to prioritize urgent referrals meant that more individuals with urological cancer were seen within 2 weeks, but at additional cost to the NHS. If used as a triage and prioritization tool, the PinPoint test shortens wait times for referred individuals and is cost saving. More data on the impact of short-term delays to diagnosis on health-related quality of life is needed.
State Medicine , Urologic Neoplasms , Male , Humans , Cost-Benefit Analysis , Quality of Life , Urologic Neoplasms/diagnosis , Referral and Consultation
BACKGROUND: The early detection and diagnosis of cancer to reduce avoidable mortality and morbidity is a challenging task in primary health care. There is a growing evidence base on how to enable earlier cancer diagnosis, but well-recognised gaps and delays exist around the translation of new research findings into routine clinical practice. Implementation research aims to accelerate the uptake of evidence by health care systems and professionals. We aimed to identify priorities for implementation research in early cancer diagnosis in primary care. METHODS: We used a RAND/UCLA modified Delphi consensus process to identify and rank research priorities. We asked primary care physicians, patients and researchers to complete an online survey suggesting priorities for implementation research in cancer detection and diagnosis. We summarised and presented these suggestions to an 11-member consensus panel comprising nine primary care physicians and two patients. Panellists independently rated the importance of suggestions on a 1-9 scale (9 = very high priority; 1 = very low priority) before and after a structured group discussion. We ranked suggestions using median ratings. RESULTS: We received a total of 115 suggested priorities for implementation research from 32 survey respondents (including 16 primary care professionals, 11 researchers, and 4 patient and public representatives; 88% of respondents were UK-based). After removing duplicates and ineligible suggestions, we presented 37 suggestions grouped within 17 categories to the consensus panel. Following two rounds of rating, 27 suggestions were highly supported (median rating 7-9). The most highly rated suggestions concerned diagnostic support (e.g., access to imaging) interventions (e.g., professional or patient education), organisation of the delivery of care (e.g., communication within and between teams) and understanding variations in care and outcomes. CONCLUSIONS: We have identified a set of priorities for implementation research on the early diagnosis of cancer, ranked in importance by primary care physicians and patients. We suggest that researchers and research funders consider these in directing further efforts and resources to improve population outcomes.
Neoplasms , Humans , Consensus , Neoplasms/diagnosis , Palliative Care , Delphi Technique , Primary Health Care
BACKGROUND: Global annual cancer incidence is forecast to rise to 27.5 M by 2040, a 62% increase from 2018. For most cancers, prevention and early detection are the most effective ways of reducing mortality. This study maps trials in cancer screening, prevention, and early diagnosis (SPED) to identify areas of unmet need and highlight research priorities. METHODS: A systematic mapping review was conducted to evaluate all clinical trials focused on cancer SPED, irrespective of tumour type. The National Cancer Research Institute (NCRI) portfolio, EMBASE, PubMed and Medline were searched for relevant papers published between 01/01/2007 and 01/04/2020. References were exported into Covidence software and double-screened. Data were extracted and mapped according to tumour site, geographical location, and intervention type. RESULTS: One hundred seventeen thousand seven hundred one abstracts were screened, 5157 full texts reviewed, and 2888 studies included. 1184 (52%) trials focussed on screening, 554 (24%) prevention, 442 (20%) early diagnosis, and 85 (4%) a combination. Colorectal, breast, and cervical cancer comprised 61% of all studies compared with 6.4% in lung and 1.8% in liver cancer. The latter two are responsible for 26.3% of global cancer deaths compared with 19.3% for the former three. Number of studies varied markedly according to geographical location; 88% were based in North America, Europe, or Asia. CONCLUSIONS: This study shows clear disparities in the volume of research conducted across different tumour types and according to geographical location. These findings will help drive future research effort so that resources can be directed towards major challenges in cancer SPED.
Liver Neoplasms , Uterine Cervical Neoplasms , Female , Humans , Early Detection of Cancer , Asia , Breast
Introduction: COPD is underdiagnosed, and measurement of spirometry alongside low-dose computed tomography (LDCT) screening for lung cancer is one strategy to increase earlier diagnosis of this disease. Methods: Ever-smokers at high risk of lung cancer were invited to the Yorkshire Lung Screening Trial for a lung health check (LHC) comprising LDCT screening, pre-bronchodilator spirometry and a smoking cessation service. In this cross-sectional study we present data on participant demographics, respiratory symptoms, lung function, emphysema on imaging and both self-reported and primary care diagnoses of COPD. Multivariable logistic regression analysis identified factors associated with possible underdiagnosis and misdiagnosis of COPD in this population, with airflow obstruction defined as forced expiratory volume in 1â s/forced vital capacity ratio <0.70. Results: Out of 3920 LHC attendees undergoing spirometry, 17% had undiagnosed airflow obstruction with respiratory symptoms, representing potentially undiagnosed COPD. Compared to those with a primary care COPD code, this population had milder symptoms, better lung function and were more likely to be current smokers (p≤0.001 for all comparisons). Out of 836 attendees with a primary care COPD code who underwent spirometry, 19% did not have airflow obstruction, potentially representing misdiagnosed COPD, although symptom burden was high. Discussion: Spirometry offered alongside LDCT screening can potentially identify cases of undiagnosed and misdiagnosed COPD. Future research should assess the downstream impact of these findings to determine whether any meaningful changes to treatment and outcomes occur, and to assess the impact on co-delivering spirometry on other parameters of LDCT screening performance such as participation and adherence. Additionally, work is needed to better understand the aetiology of respiratory symptoms in those with misdiagnosed COPD, to ensure that this highly symptomatic group receive evidence-based interventions.
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).
Multiple myeloma (MM) patients risk diagnostic delays and irreversible organ damage. In those with newly diagnosed myeloma, we explored the presenting symptoms to identify early signals of MM and their relationships to organ damage. The symptoms were recorded in patients' own words at diagnosis and included diagnostic time intervals. Those seen by a haematologist >6 months prior to MM diagnosis were classified as precursor disease (PD). Most (962/977) patients provided data. Back pain (38%), other pain (31%) and systemic symptoms (28%) predominated. Patients rarely complain of 'bone pain', simply 'pain'. Vertebral fractures are under-recognised as pathological and are the predominant irreversible organ damage (27% of patients), impacting the performance status (PS) and associated with back pain (odds ratio (OR) 6.14 [CI 4.47-8.44]), bone disease (OR 3.71 [CI 1.88-7.32]) and age >65 years (OR 1.58 [CI 1.15-2.17]). Renal failure is less frequent and associated with gastrointestinal symptoms (OR 2.23 [CI1.28-3.91]), age >65 years (OR 2.14 [CI1.28-3.91]) and absence of back pain (OR 0.44 [CI 0.29-0.67]). Patients with known PD (n = 149) had fewer vertebral fractures (p = 0.001), fewer adverse features (p = 0.001), less decline in PS (p = 0.001) and a lower stage (p = 0.04) than 813 with de novo MM. Our data suggest subgroups suitable for trials of 'symptom-directed' screening: those with back pain, unexplained pain, a general decline in health or low-impact vertebral compression fractures.
BACKGROUND: Cancer is predominantly a disease of older adults. To date there has been little research on the experiences of older adults or their views on the diagnostic pathway. AIM: To gain an improved understanding of the views and experiences of older adults on all aspects of cancer investigation. DESIGN AND SETTING: This was a qualitative study using semi-structured interviews with patients aged ≥70 years. Patients were recruited from primary care in West Yorkshire, UK. METHOD: Data were analysed using a thematic framework analysis. RESULTS: The themes identified in participants' accounts included the patients' process of decision making, the value of having a diagnosis, the patients' experience of cancer investigations, and the impact of the COVID-19 pandemic on the diagnostic pathway. Older adults in this study indicated a clear preference for having clarity on the cause of symptoms and the diagnosis, even in the face of unpleasant investigations. Patients suggested they wanted to be involved in the decision process. CONCLUSION: Older adults who present to primary care with symptoms suggestive of cancer may accept diagnostic testing solely for the benefit of knowing the diagnosis. There was a clear patient preference that referrals and investigations for cancer symptoms should not be deferred or delayed based on age or subjective assessments of frailty. Shared decision making and being involved in the decision-making process are important to patients, regardless of age.
COVID-19 , Neoplasms , Humans , Aged , Pandemics , COVID-19/diagnosis , Neoplasms/diagnosis , Qualitative Research , Primary Health Care , Decision Making
BACKGROUND: Screening programmes utilising blood-based multi-cancer early detection (MCED) tests, which can detect a shared cancer signal from any site in the body with a single, low false-positive rate, could reduce cancer burden through early diagnosis. METHODS: A natural history ('interception') model of cancer was previously used to characterise potential benefits of MCED screening (based on published performance of an MCED test). We built upon this using a two-population survival model to account for an increased risk of death from cfDNA-detectable cancers relative to cfDNA-non-detectable cancers. We developed another model allowing some cancers to metastasise directly from stage I, bypassing intermediate tumour stages. We used incidence and survival-by-stage data from the National Cancer Registration and Analysis Service in England to estimate longer-term benefits to a cohort screened between ages 50-79 years. RESULTS: Estimated late-stage and mortality reductions were robust to a range of assumptions. With the least favourable dwell (sojourn) time and cfDNA status hazard ratio assumptions, we estimated, among 100,000 screened individuals, 67 (17%) fewer cancer deaths per year corresponding to 2029 fewer deaths in those screened between ages 50-79 years. CONCLUSION: Realising the potential benefits of MCED tests could substantially reduce late-stage cancer diagnoses and mortality.
Early Detection of Cancer , Neoplasms , Humans , England/epidemiology , Neoplasms/diagnosis , Mass Screening
OBJECTIVE: Lung cancer is the most common cause of cancer-related death in the USA. While most patients are diagnosed following symptomatic presentation, no studies have compared symptoms and physical examination signs at or prior to diagnosis from electronic health records (EHRs) in the USA. We aimed to identify symptoms and signs in patients prior to diagnosis in EHR data. DESIGN: Case-control study. SETTING: Ambulatory care clinics at a large tertiary care academic health centre in the USA. PARTICIPANTS, OUTCOMES: We studied 698 primary lung cancer cases in adults diagnosed between 1 January 2012 and 31 December 2019, and 6841 controls matched by age, sex, smoking status and type of clinic. Coded and free-text data from the EHR were extracted from 2 years prior to diagnosis date for cases and index date for controls. Univariate and multivariable conditional logistic regression were used to identify symptoms and signs associated with lung cancer at time of diagnosis, and 1, 3, 6 and 12 months before the diagnosis/index dates. RESULTS: Eleven symptoms and signs recorded during the study period were associated with a significantly higher chance of being a lung cancer case in multivariable analyses. Of these, seven were significantly associated with lung cancer 6 months prior to diagnosis: haemoptysis (OR 3.2, 95% CI 1.9 to 5.3), cough (OR 3.1, 95% CI 2.4 to 4.0), chest crackles or wheeze (OR 3.1, 95% CI 2.3 to 4.1), bone pain (OR 2.7, 95% CI 2.1 to 3.6), back pain (OR 2.5, 95% CI 1.9 to 3.2), weight loss (OR 2.1, 95% CI 1.5 to 2.8) and fatigue (OR 1.6, 95% CI 1.3 to 2.1). CONCLUSIONS: Patients diagnosed with lung cancer appear to have symptoms and signs recorded in the EHR that distinguish them from similar matched patients in ambulatory care, often 6 months or more before diagnosis. These findings suggest opportunities to improve the diagnostic process for lung cancer.
Electronic Health Records , Lung Neoplasms , Adult , Humans , Case-Control Studies , Tertiary Care Centers , Lung Neoplasms/diagnosis , Ambulatory Care
A comprehensive evaluation of the total burden of morbidity endured by cancer survivors remains unavailable. This study quantified the burden of 144 health conditions and critical care admissions across 26 adult cancers and treatment modalities in 243,767 adults. By age 60, top conditions ranked by fold difference (cumulative burden in survivors divided by cumulative burden in controls) were haematology, immunology/infection and pulmonary conditions. Patients who had all three forms of treatment (chemotherapy, radiotherapy and surgery) experienced a high cumulative burden of late morbidities compared with patients who received radiotherapy alone. The top five cancers with the highest cumulative burden of critical care admissions by age 60 were bone (12.4 events per 100 individuals [CI: 11.6-13.1]), brain (9.0 [7.5-10.5]), spinal cord and nervous system (7.2 [6.7-7.8]), testis (6.7 [4.9-8.4]) and Hodgkin lymphoma (4.4 [3.6-5.1]). Conditions that were associated with high excess years-of-life-lost were haematological conditions (9.6 years), pulmonary conditions (8.6 years) and immunological conditions or infections (7.8 years). As the population of cancer survivors continues to grow, our results indicate that it is important to tackle long-term health consequences through enacting data-driven policies.
Cancer Survivors , Hodgkin Disease , Male , Adult , Humans , Middle Aged , Mortality, Premature , Morbidity , Hospitalization , Risk Factors
Myeloma is one of the hardest cancers to diagnose in primary care due to its rarity and non-specific symptoms. A rate-limiting step in diagnosing myeloma is the clinician considering myeloma and initiating appropriate investigations. We developed and internally validated a risk prediction model to identify those with a high risk of having undiagnosed myeloma based on results from routine blood tests taken for other reasons. A case-control study, based on 367 myeloma cases and 1488 age- and sex-matched controls, was used to develop a risk prediction model including results from 15 blood tests. The model had excellent discrimination (C-statistic 0.85 (95%CI 0.83, 0.89)) and good calibration (calibration slope 0.87 (95%CI 0.75, 0.90)). At a prevalence of 15 per 100,000 population and a probability threshold of 0.4, approximately 600 patients would need additional reflex testing to detect one case. We showed that it is possible to combine signals and abnormalities from several routine blood test parameters to identify individuals at high-risk of having undiagnosed myeloma who may benefit from additional reflex testing. Further work is needed to explore the full potential of such a strategy, including whether it is clinically useful and cost-effective and how to make it ethically acceptable.
As many countries experience population aging, patients with cancer are becoming older and have more preexisting comorbidities, which include prevalent, age-related, chronic conditions such as dementia. People living with dementia (PLWD) are vulnerable to health disparities, and dementia has high potential to complicate and adversely affect care and outcomes across the cancer trajectory. This report offers an overview of dementia and its prevalence among patients with cancer and a summary of the research literature examining cancer care for PLWD. The reviewed research indicates that PLWD are more likely to have cancer diagnosed at an advanced stage, receive no or less extensive cancer treatment, and have poorer survival after a cancer diagnosis. These cancer disparities do not necessarily signify inappropriately later diagnosis or lower treatment of people with dementia as a group, and they are arguably less feasible and appropriate targets for care optimization. The reviewed research indicates that PLWD also have an increased risk of cancer-related emergency presentations, lower quality processes of cancer-related decision making, accessibility-related barriers to cancer investigations and treatment, higher experienced treatment burden and higher caregiver burden for families, and undertreated cancer-related pain. The authors propose that optimal cancer care for PLWD should focus on proactively minimizing these risk areas and thus must be highly person-centered, with holistic decision making, individualized reasonable adjustments to practice, and strong inclusion and support of family carers. Comprehensive recommendations are made for clinical practice and future research to help clinicians and providers deliver best and equitable cancer care for PLWD and their families.
Dementia , Neoplasms , Humans , Dementia/complications , Dementia/diagnosis , Dementia/therapy , Caregivers , Neoplasms/complications , Neoplasms/therapy
OBJECTIVES: A growing body of evidence suggests longer time between symptom onset and start of treatment affects breast cancer prognosis. To explore this association, the International Cancer Benchmarking Partnership Module 4 examined differences in breast cancer diagnostic pathways in 10 jurisdictions across Australia, Canada, Denmark, Norway, Sweden and the UK. SETTING: Primary care in 10 jurisdictions. PARTICIPANT: Data were collated from 3471 women aged >40 diagnosed for the first time with breast cancer and surveyed between 2013 and 2015. Data were supplemented by feedback from their primary care physicians (PCPs), cancer treatment specialists and available registry data. PRIMARY AND SECONDARY OUTCOME MEASURES: Patient, primary care, diagnostic and treatment intervals. RESULTS: Overall, 56% of women reported symptoms to primary care, with 66% first noticing lumps or breast changes. PCPs reported 77% presented with symptoms, of whom 81% were urgently referred with suspicion of cancer (ranging from 62% to 92%; Norway and Victoria). Ranges for median patient, primary care and diagnostic intervals (days) for symptomatic patients were 3-29 (Denmark and Sweden), 0-20 (seven jurisdictions and Ontario) and 8-29 (Denmark and Wales). Ranges for median treatment and total intervals (days) for all patients were 15-39 (Norway, Victoria and Manitoba) and 4-78 days (Sweden, Victoria and Ontario). The 10% longest waits ranged between 101 and 209 days (Sweden and Ontario). CONCLUSIONS: Large international differences in breast cancer diagnostic pathways exist, suggesting some jurisdictions develop more effective strategies to optimise pathways and reduce time intervals. Targeted awareness interventions could also facilitate more timely diagnosis of breast cancer.
Benchmarking , Breast Neoplasms , Humans , Female , Breast Neoplasms/diagnosis , Cohort Studies , Routinely Collected Health Data , Ontario , Surveys and Questionnaires , Victoria
The diagnosis of lung cancer in ambulatory settings is often challenging due to non-specific clinical presentation, but there are currently no clinical quality measures (CQMs) in the United States used to identify areas for practice improvement in diagnosis. We describe the pre-diagnostic time intervals among a retrospective cohort of 711 patients identified with primary lung cancer from 2012-2019 from ambulatory care clinics in Seattle, Washington USA. Electronic health record data were extracted for two years prior to diagnosis, and Natural Language Processing (NLP) applied to identify symptoms/signs from free text clinical fields. Time points were defined for initial symptomatic presentation, chest imaging, specialist consultation, diagnostic confirmation, and treatment initiation. Median and interquartile ranges (IQR) were calculated for intervals spanning these time points. The mean age of the cohort was 67.3 years, 54.1% had Stage III or IV disease and the majority were diagnosed after clinical presentation (94.5%) rather than screening (5.5%). Median intervals from first recorded symptoms/signs to diagnosis was 570 days (IQR 273-691), from chest CT or chest X-ray imaging to diagnosis 43 days (IQR 11-240), specialist consultation to diagnosis 72 days (IQR 13-456), and from diagnosis to treatment initiation 7 days (IQR 0-36). Symptoms/signs associated with lung cancer can be identified over a year prior to diagnosis using NLP, highlighting the need for CQMs to improve timeliness of diagnosis.
Over 60% of cancer diagnoses in the UK are in patients aged 65 and over. Cancer diagnosis and treatment in older adults is complicated by the presence of frailty, which is associated with lower survival rates and poorer quality of life. This population-based cohort study used a longitudinal database to calculate the time between presentation to primary care with a symptom suspicious of cancer and a confirmed cancer diagnosis for 7460 patients in the Bradford District. Individual frailty scores were calculated using the electronic frailty index (eFI) and categorised by severity. The median time from symptomatic presentation to cancer diagnosis for all patients was 48 days (IQR 21-142). 23% of the cohort had some degree of frailty. After adjustment for potential confounders, mild frailty added 7 days (95% CI 3-11), moderate frailty 23 days (95% CI 4-42) and severe frailty 11 days (95% CI -27-48) to the median time to diagnosis compared to not frail patients. Our findings support use of the eFI in primary care to identify and address patient, healthcare and system factors that may contribute to diagnostic delay. We recommend further research to explore patient and clinician factors when investigating cancer in frail patients.
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.
