Colorectal Cancer In Primary Care Research Articles

71629 - Comparison of machine learning tools for identifying non-metastatic and metastatic colorectal cancer in primary care

71629 - Comparison of machine learning tools for identifying non-metastatic and metastatic colorectal cancer in primary care

Early detection of colorectal cancer by leveraging Dutch primary care consultation notes with free text embeddings

We aimed to assess the added predictive performance that free-text Dutch consultation notes provide in detecting colorectal cancer in primary care, in comparison to currently used models. We developed, evaluated and compared three prediction models for colorectal cancer (CRC) in a large primary care database with 60,641 patients. The prediction model with both known predictive features and free-text data (with TabTxt AUROC: 0.823) performs statistically significantly better (p < 0.05) than the other two models with only tabular (as used nowadays) and text data, respectively (AUROC Tab: 0.767; Txt: 0.797). The specificity of the two models that use demographics and known CRC features (with specificity Tab: 0.321; TabTxt: 0.335) are higher than that of the model with only free-text (specificity Txt: 0.234). The Txt and, to a lesser degree, TabTxt model are well calibrated, while the Tab model shows slight underprediction at both tails. As expected with an outcome prevalence below 0.01, all models show much uncalibrated predictions in the extreme upper tail (top 1%). Free-text consultation notes show promising results to improve the predictive performance over established prediction models that only use structured features. Clinical future implications for our CRC use case include that such improvement may help lowering the number of referrals for suspected CRC to medical specialists.

A machine learning tool for identifying non-metastatic colorectal cancer in primary care

BackgroundPrimary health care (PHC) is often the first point of contact when diagnosing colorectal cancer (CRC). Human limitations in processing large amounts of information warrant the use of machine learning as a diagnostic prediction tool for CRC. AimTo develop a predictive model for identifying non-metastatic CRC (NMCRC) among PHC patients using diagnostic data analysed with machine learning. Design and settingA case–control study containing data on PHC visits for 542 patients >18 years old diagnosed with NMCRC in the Västra Götaland Region, Sweden, during 2011, and 2,139 matched controls. MethodStochastic gradient boosting (SGB) was used to construct a model for predicting the presence of NMCRC based on diagnostic codes from PHC consultations during the year before the date of cancer diagnosis and the total number of consultations. Variables with a normalised relative influence (NRI) >1% were considered having an important contribution to the model. Risks of having NMCRC were calculated using odds ratios of marginal effects. ResultsOf the 361 variables used as predictors in the stochastic gradient boosting model, 184 had non-zero influence, with 16 variables having NRI >1% and a combined NRI of 63.3%. Variables representing anaemia and bleeding had a combined NRI of 27.6%. The model had a sensitivity of 73.3% and a specificity of 83.5%. Change in bowel habit had the highest odds ratios of marginal effects at 28.8. ConclusionMachine learning is useful for identifying variables of importance for predicting NMCRC in PHC. Malignant diagnoses may be hidden behind benign symptoms such as haemorrhoids.

Validation of a diagnostic prediction tool for colorectal cancer: a case-control replication study.

Early detection of colorectal cancer (CRC) is crucial for survival. Primary care, the first point of contact in most cases, needs supportive risk assessment tools. We aimed to replicate the Swedish Colorectal Cancer Risk Assessment Tool (SCCRAT) for non-metastatic CRC in primary care and examine if risk factor patterns depend on sex and age. 2,920 adults diagnosed with non-metastatic CRC during the years 2015-2019 after having visited a general practitioner the year before the diagnosis were selected from the Swedish Cancer Register and matched with 11,628 controls, using the same inclusion criteria except for the CRC diagnosis. Diagnostic codes from primary care consultations were collected from a regional health care database. Positive predictive values (PPVs) were estimated for the same 5 symptoms and combinations thereof as in the baseline study. The results for patients aged ≥50 years old in the present study were consistent with the results of the SCCRAT study. All symptoms and combinations thereof with a PPV >5% in the present study had a PPV >5% in the baseline study. The combination of bleeding with abdominal pain (PPV 9.9%) and bleeding with change in bowel habit (PPV 7.8%) were the highest observed PPVs in both studies. Similar risk patterns were seen for all ages and when men and women were studied separately. This external validation of the SCCRAT for non-metastatic CRC in primary care replicated the baseline study successfully and identified patients at high risk for CRC.

A novel blood based triage test for colorectal cancer in primary care: a pilot study.

The majority of colorectal cancers (CRCs) are detected after symptomatic presentation to primary care. Given the shared symptoms of CRC and benign disorders, it is challenging to manage the risk of missed diagnosis. Colonoscopy resources cannot keep pace with increasing demand. There is a pressing need for access to simple triage tools in primary care to help prioritise patients for referral. To evaluate the performance of a novel spectroscopy-based CRC blood test in primary care. Mixed-methods pilot study of test performance and GP focus group discussions in South Wales. Patients on the urgent suspected cancer (USC) pathway were recruited for the Raman spectroscopy (RS) test coupled to machine learning classification ('Raman-CRC') to identify CRC within the referred population. Qualitative focus group work evaluated the acceptability of the test in primary care by thematic analysis of focus group theorising. A total of 532 patients aged ≥50 years referred on the USC pathway were recruited from 27 GP practices. Twenty-nine patients (5.0%) were diagnosed with CRC. Raman-CRC identified CRC with sensitivity 95.7%, specificity 69.3% with area under curve (AUC) of 0.80 compared with colonoscopy as the reference test (248 patients). Stage I and II cancers were detected with 78.6% sensitivity. Focus group themes underlined the convenience of a blood test for the patient and the test's value as a risk assessment tool in primary care. The findings support this novel, non-invasive, blood-based method to prioritise those patients most likely to have CRC. Raman-CRC may accelerate access to diagnosis with potential to improve cancer outcomes.

Faecal immunochemical tests safely enhance rational use of resources during the assessment of suspected symptomatic colorectal cancer in primary care: systematic review and meta-analysis

ObjectiveImplementation of faecal immunochemical tests (FIT) as a triage test in primary healthcare may improve the efficiency of referrals without missing cases of colorectal cancer (CRC). We aim to summarise...

Recognising Colorectal Cancer in Primary Care.

Colorectal cancer (CRC) is the third most common cancer worldwide. Primary care professionals can play an important role in both prevention and early detection of CRC. Most CRCs are attributed to modifiable lifestyle factors, which can be addressed within primary care, and promotion of population-based screening programmes can aid early cancer detection in asymptomatic patients. Primary care professionals have a vital role in clinically assessing patients presenting with symptoms that may indicate cancer, as most patients with CRC first present with symptoms. These assessments are often challenging—many of the symptoms of CRC are non-specific and commonly occur in patients presenting with non-malignant disease. The range of options for investigating symptomatic patients in primary care is rapidly growing. Simple tests, such as faecal immunochemical testing (FIT), are now being used to guide decisions around referral for more invasive tests, such as colonoscopy, while direct access to specialist investigations is also becoming more common. Clinical decision support tools (CDSTs) which calculate cancer risk based on symptomatology, patient characteristics and test results can provide an additional resource to guide decisions on further investigation. This article explores the challenges of CRC prevention and detection from the primary care perspective, discusses current evidence-based approaches for CRC detection used in primary care (with examples from UK guidelines), and highlights emerging research which may likely alter practice in the future.

Diagnostic performance of a faecal immunochemical test for patients with low-risk symptoms of colorectal cancer in primary care: an evaluation in the South West of England

BackgroundThe faecal immunochemical test (FIT) was introduced to triage patients with low-risk symptoms of possible colorectal cancer in English primary care in 2017, underpinned by little primary care evidence.MethodsAll healthcare providers in the South West of England (population 4 million) participated in this evaluation. 3890 patients aged ≥50 years presenting in primary care with low-risk symptoms of colorectal cancer had a FIT from 01/06/2018 to 31/12/2018. A threshold of 10 μg Hb/g faeces defined a positive test.ResultsSix hundred and eighteen (15.9%) patients tested positive; 458 (74.1%) had an urgent referral to specialist lower gastrointestinal (GI) services within three months. Forty-three were diagnosed with colorectal cancer within 12 months. 3272 tested negative; 324 (9.9%) had an urgent referral within three months. Eight were diagnosed with colorectal cancer within 12 months. Positive predictive value was 7.0% (95% CI 5.1–9.3%). Negative predictive value was 99.8% (CI 99.5–99.9%). Sensitivity was 84.3% (CI 71.4–93.0%), specificity 85.0% (CI 83.8–86.1%). The area under the ROC curve was 0.92 (CI 0.86–0.96). A threshold of 37 μg Hb/g faeces would identify patients with an individual 3% risk of cancer.ConclusionsFIT performs exceptionally well to triage patients with low-risk symptoms of colorectal cancer in primary care; a higher threshold may be appropriate in the wake of the COVID-19 crisis.

Development, validation and effectiveness of diagnostic prediction tools for colorectal cancer in primary care: a systematic review

BackgroundTools based on diagnostic prediction models are available to help general practitioners (GP) diagnose colorectal cancer. It is unclear how well they perform and whether they lead to increased or quicker diagnoses and ultimately impact on patient quality of life and/or survival. The aim of this systematic review is to evaluate the development, validation, effectiveness, and cost-effectiveness, of cancer diagnostic tools for colorectal cancer in primary care.MethodsElectronic databases including Medline and Web of Science were searched in May 2017 (updated October 2019). Two reviewers independently screened titles, abstracts and full-texts. Studies were included if they reported the development, validation or accuracy of a prediction model, or assessed the effectiveness or cost-effectiveness of diagnostic tools based on prediction models to aid GP decision-making for symptomatic patients presenting with features potentially indicative of colorectal cancer. Data extraction and risk of bias were completed by one reviewer and checked by a second. A narrative synthesis was conducted.ResultsEleven thousand one hundred thirteen records were screened and 23 studies met the inclusion criteria. Twenty-studies reported on the development, validation and/or accuracy of 13 prediction models: eight for colorectal cancer, five for cancer areas/types that include colorectal cancer. The Qcancer models were generally the best performing.Three impact studies met the inclusion criteria. Two (an RCT and a pre-post study) assessed tools based on the RAT prediction model. The third study looked at the impact of GP practices having access to RAT or Qcancer.Although the pre-post study reported a positive impact of the tools on outcomes, the results of the RCT and cross-sectional survey found no evidence that use of, or access to, the tools was associated with better outcomes. No study evaluated cost effectiveness.ConclusionsMany prediction models have been developed but none have been fully validated. Evidence demonstrating improved patient outcome of introducing the tools is the main deficiency and is essential given the imperfect classification achieved by all tools. This need is emphasised by the equivocal results of the small number of impact studies done so far.

Diagnosing colorectal cancer in primary care: cohort study in Sweden of qualitative faecal immunochemical tests, haemoglobin levels, and platelet counts.

BackgroundColorectal cancer (CRC) diagnostics are challenging in primary care and reliable diagnostic aids are desired. Qualitative faecal immunochemical tests (FITs) have been used for suspected CRC in Sweden since the mid-2000s, but evidence regarding their effectiveness is scarce. Anaemia and thrombocytosis are both associated with CRC.AimTo evaluate the usefulness of qualitative FITs requested for symptomatic patients in primary care, alone and combined with findings of anaemia and thrombocytosis, in the diagnosis of CRC.Design and settingA population-based cohort study using electronic health records and data from the Swedish Cancer Register, covering five Swedish regions.MethodPatients aged ≥18 years in the five regions who had provided FITs requested by primary care practitioners from 1 January 2015 to 31 December 2015 were identified. FIT and blood-count data were registered and all CRC diagnoses made within 2 years were retrieved. Diagnostic measurements were calculated.ResultsIn total, 15 789 patients provided FITs (four different brands); of these patients, 304 were later diagnosed with CRC. Haemoglobin levels were available for 13 863 patients, and platelet counts for 10 973 patients. Calculated for the different FIT brands only, the sensitivities for CRC were 81.6%–100%; specificities 65.7%–79.5%; positive predictive values 4.7%–8.1%; and negative predictive values 99.5%–100%. Calculated for the finding of either a positive FIT or anaemia, the sensitivities increased to 88.9–100%. Adding thrombocytosis did not further increase the diagnostic performance.ConclusionQualitative FITs requested in primary care seem to be useful as rule-in tests for referral when CRC is suspected. A negative FIT and no anaemia indicate a low risk of CRC.

Faecal immunochemical test to triage patients with abdominal symptoms for suspected colorectal cancer in primary care: review of international use and guidelines.

Recently, faecal immunochemical tests (FITs) have been introduced for investigation of primary care patients with low-risk symptoms of colorectal cancer (CRC), but recommendations vary across the world. This systematic review of clinical practice guidelines aimed to determine how FITs are used in symptomatic primary care patients and the underpinning evidence for these guidelines. MEDLINE, Embase and TRIP databases were systematically searched, from 1 November 2008 to 1 November 2018 for guidelines on the assessment of patients with symptoms suggestive of CRC. Known guideline databases, websites and references of related literature were searched. The following questions were addressed: (i) which countries use FIT for symptomatic primary care patients; (ii) in which populations is FIT used; (iii) what is the cut-off level used for haemoglobin in the faeces (FIT) and (iv) on what evidence are FIT recommendations based. The search yielded 2433 publications; 25 covered initial diagnostic assessment of patients with symptoms of CRC in 15 countries (Asia, n = 1; Europe, n = 13; Oceania, n = 4; North America, n = 5; and South America, n = 2). In three countries (Australia, Spain and the UK), FIT was recommended for patients with abdominal symptoms, unexplained weight loss, change in bowel habit or anaemia despite a low level of evidence in the symptomatic primary care patient population. Few countries recommend FITs in symptomatic patients in primary care either because of limited evidence or because symptomatic patients are directly referred to secondary care without triage. These results demonstrate a clear need for research on FIT in the symptomatic primary care population.

Data mining-based model and risk prediction of colorectal cancer by using secondary health data: A systematic review.

ObjectivePrevention and early detection of colorectal cancer (CRC) can increase the chances of successful treatment and reduce burden. Various data mining technologies have been utilized to strengthen the early detection of CRC in primary care. Evidence synthesis on the model’s effectiveness is scant. This systematic review synthesizes studies that examine the effect of data mining on improving risk prediction of CRC.MethodsThe PRISMA framework guided the conduct of this study. We obtained papers via PubMed, Cochrane Library, EMBASE and Google Scholar. Quality appraisal was performed using Downs and Black’s quality checklist. To evaluate the performance of included models, the values of specificity and sensitivity were comparted, the values of area under the curve (AUC) were plotted, and the median of overall AUC of included studies was computed.ResultsA total of 316 studies were reviewed for full text. Seven articles were included. Included studies implement techniques including artificial neural networks, Bayesian networks and decision trees. Six articles reported the overall model accuracy. Overall, the median AUC is 0.8243 [interquartile range (IQR): 0.8050−0.8886]. In the two articles that reported comparison results with traditional models, the data mining method performed better than the traditional models, with the best AUC improvement of 10.7%.ConclusionsThe adoption of data mining technologies for CRC detection is at an early stage. Limited numbers of included articles and heterogeneity of those studies implied that more rigorous research is expected to further investigate the techniques’ effects.

Potential for Reducing Time to Referral for Colorectal Cancer Patients in Primary Care.

An optimal diagnostic process in primary care is pivotal for reducing cancer-related disease burden. This study aims to explore reasons for long times to referral for Dutch colorectal cancer (CRC) patients in primary care. A retrospective cohort study of anonymized free-text primary care records from the Julius General Practitioners' Network database, linked to the Netherlands Cancer Registry. Patients with a confirmed CRC diagnosis from 2007 through 2011 that symptomatically presented in primary care were included. Median time and interquartile ranges from presentation in primary care to referral were calculated for multiple patient and presentation characteristics. Associations of these characteristics with long time to referral (75th percentile was ≥59 days) were examined with log-binomial regression analyses. Routes to referral of patients with the longest times to referral were explored using thematic free-text analyses (90th percentile at ≥219 days). Among the 309 people with CRC, patients who were female, did not have a registered family history, had a history of malignancy, lacked alarm symptoms at presentation, or had hemorrhoids at physical examination were at risk for longer time to referral in univariable analyses (longer median durations and/or univariable association with the 75th percentile). Only presentation without alarm symptoms showed a statistically significant association with long duration (75th percentile) in multivariable analysis (relative risk = 1.7; 95% CI, 1.1-2.6). Thematic exploration of the diagnostic routes to referral of patients with the longest durations (90th percentile) showed 2 dominating themes: "alternative working diagnosis" and "suboptimal diagnostic strategies," and included the sub-themes "omitting to reconsider an initial diagnosis" and "lacking follow-up." Long time to referral for CRC in primary care is mainly related to low cancer suspicion. There is potential for reducing the longest times to referral for patients with CRC in primary care, with earlier reconsideration of the initial hypothesis and implementation of strict follow-up consultations.

Faecal immunochemical tests for the diagnosis of symptomatic colorectal cancer in primary care: the benefit of more than one sample

Objective: Faecal immunochemical tests (FITs) are used to screen for colorectal cancer (CRC) and as diagnostic aids in symptomatic patients. However, the number of samples per FIT varies. It is unclear if there is any advantage to analyse multiple-sample FITs in symptomatic patients.Design and setting: This is a post hoc analysis of a retrospective study that included all cases of CRC and adenomas with high-grade dysplasia (HGD) between 2005 and 2009 in the county of Jämtland, Sweden.Subjects: All patients with CRC and adenomas with HGD that initially presented with symptoms to primary care and delivered FITs.Main outcome measure: The likelihood of a positive FIT in cases of CRC and adenomas with HGD; when analysing one, two or three samples.Results: Of 195 patients, 160 delivered three-sample FITs. Using the 139 cases in which at least one sample was positive, the likelihood of detecting a positive sample upon analysis of only one of the three samples was 0.91 (95% CI: 0.85–0.95), indicating that 13 positive cases may have been missed.Conclusion: Use of a one-sample FIT instead of a three-sample FIT as a diagnostic aid may result in the missing of one tenth of symptomatic CRCs and adenomas with HGD.

Volume-outcome relationship for colorectal cancer in primary care: a prospective cohort study

PurposeHigher caseloads are associated with better outcomes for many conditions treated in secondary and tertiary care settings, including colorectal cancer (CRC). There is little known whether such volume-outcome relationship exist in primary care settings. The purpose of this paper is to examine general practitioner (GP) CRC-specific caseload for possible associations with referral pathways, disease stage and CRC patients’ overall survival.Design/methodology/approachThe paper retrospectively analyses a prospectively maintained CRC database for 2009-2014 in a single district hospital providing bowel cancer screening and tertiary rectal cancer services.FindingsOf 1,145 CRC patients, 937 (81.8 per cent) were diagnosed as symptomatic cancers. In total, 210 GPs from 44 practices were stratified according to their CRC caseload over the study period into low volume (LV, 1-4); medium volume (MV, 5-7); and high volume (HV, 8-21 cases). Emergency presentation (LV: 49/287 (17.1 per cent); MV: 75/264 (28.4 per cent); HV: 105/386 (27.2 per cent); p=0.007) and advanced disease at presentation (LV: 84/287 (29.3 per cent); MV: 94/264 (35.6 per cent); HV: 144/386 (37.3 per cent); p=0.034) was more common amongst HV GPs. Three-year mortality risk was significantly higher for HV GPs (MV: (hazard ratio) HR 1.185 (confidence interval=0.897-1.566), p=0.231, and HV: HR 1.366 (CI=1.061-1.759), p=0.016), but adjustment for emergency presentation and advanced disease largely accounted for this difference. There was some evidence that HV GPs used elective cancer pathways less frequently (LV: 166/287 (57.8 per cent); MV: 130/264 (49.2 per cent); HV: 182/386 (47.2 per cent); p=0.007) and more selectively (CRC/referrals: LV: 166/2,743 (6.1 per cent); MV: 130/2,321 (5.6 per cent); HV: 182/2,508 (7.3 per cent); p=0.048).Originality/valueHigher GP CRC caseload in primary care may be associated with advanced disease and poorer survival; more work is required to determine the reasons and to develop targeted intervention at local level to improve elective referral rates.

Identification of patients with non-metastatic colorectal cancer in primary care: a case-control study.

Colorectal cancer is the third most common cancer worldwide and second most common in Europe. Despite screening, it is often diagnosed at an unfavourable stage. To identify and quantify features of non-metastatic colorectal cancer in primary care to enable earlier diagnosis by GPs. A case-control study was conducted using diagnostic codes from national and regional healthcare databases in Sweden. A total of 542 patients diagnosed with non-metastatic colorectal cancer in 2011 and 2139 matched controls were selected from the Swedish Cancer Register (SCR) and a regional healthcare database respectively. All diagnostic codes (according to ICD-10) from primary care consultations registered the year before the date of cancer diagnosis (according to the SCR) were collected from the regional database. Odds ratios were calculated for variables independently associated with non-metastatic colorectal cancer using multivariable conditional logistic regressions. Positive predictive values (PPVs) of these variables were calculated, both individually and in combination with each other. Five features were associated with colorectal cancer before diagnosis: bleeding, including rectal bleeding, melaena, and gastrointestinal bleeding (PPV 3.9%, 95% confidence interval [CI] = 2.3 to 6.3); anaemia (PPV 1.4%, 95% CI = 1.1 to 1.8); change in bowel habit (PPV 1.1%, 95% CI = 0.9 to 1.5; abdominal pain (PPV 0.9%, 95% CI = 0.7 to 1.1); and weight loss (PPV 1.0%, 95% CI = 0.3 to 3.0); all P-value <0.05. The combination of bleeding and change in bowel habit had a PPV of 13.7% (95% CI = 2.1 to 54.4); for bleeding combined with abdominal pain this was 12.2% (95% CI = 1.8 to 51.2). A risk assessment tool for non-metastatic colorectal cancer was designed. Bleeding combined with either diarrhoea, constipation, change in bowel habit, or abdominal pain are the most powerful predictors of non-metastatic colorectal cancer and should result in prompt referral for colorectal investigation.

The value of reasons for encounter in early detection of colorectal cancer

Background: Symptoms with a high predictive power for colorectal cancer (CRC) do not exist.Objective: To explore the predictive value of patients’ reason for encounter (RFE) in the two years prior to the diagnosis of CRC.Methods: A retrospective nested case-control study using prospectively collected data from electronic records in general practice over 20 years. Matching was done based on age (within two years), gender and practice. The positive likelihood ratios (LR+) and odds ratios (OR) were calculated for RFE between cases and controls in the two years before the index date.Results: We identified 184 CRC cases and matched 366 controls. Six RFEs had significant LR + and ORs for CRC, which may have high predictive power. These RFEs are part of four chapters in the International Classification of Primary Care (ICPC) that include tiredness (significant at 3–6 months prior to the diagnosis; LR+ 2.6 and OR 3.07; and from 0 to 3 months prior to the diagnosis; LR+ 2.0 and OR 2.36), anaemia (significant at three months before diagnosis; LR+ 9.8 and OR 16.54), abdominal pain, rectal bleeding and constipation (significant at 3–6 months before diagnosis; LR+ 3.0 and OR 3.33; 3 months prior to the diagnosis LR+ 8.0 and OR 18.10) and weight loss (significant at three months before diagnosis; LR+ 14.9 and OR 14.53).Conclusion: Data capture and organization in ICPC permits study of the predictive value of RFE for CRC in primary care.

Recommendations on screening for colorectal cancer in primary care.

Colorectal cancer is the second most common cause of cancer-related death in men and the third most common in women;[1][1] the lifetime probabilities of dying from colorectal cancer among men and women are 3.5% and 3.1%, respectively.[2][2] Although the burden of colorectal cancer varies across

Development and validation of a predictive model for detection of colorectal cancer in primary care by analysis of complete blood counts: a binational retrospective study.

Objective The use of risk prediction models grows as electronic medical records become widely available. Here, we develop and validate a model to identify individuals at increased risk for colorectal cancer (CRC) by analyzing blood counts, age, and sex, then determine the model’s value when used to supplement conventional screening. Materials and Methods Primary care data were collected from a cohort of 606 403 Israelis (of whom 3135 were diagnosed with CRC) and a case control UK dataset of 5061 CRC cases and 25 613 controls. The model was developed on 80% of the Israeli dataset and validated using the remaining Israeli and UK datasets. Performance was evaluated according to the area under the curve, specificity, and odds ratio at several working points. Results Using blood counts obtained 3–6 months before diagnosis, the area under the curve for detecting CRC was 0.82 ± 0.01 for the Israeli validation set. The specificity was 88 ± 2% in the Israeli validation set and 94 ± 1% in the UK dataset. Detecting 50% of CRC cases, the odds ratio was 26 ± 5 and 40 ± 6, respectively, for a false-positive rate of 0.5%. Specificity for 50% detection was 87 ± 2% a year before diagnosis and 85 ± 2% for localized cancers. When used in addition to the fecal occult blood test, our model enabled more than a 2-fold increase in CRC detection. Discussion Comparable results in 2 unrelated populations suggest that the model should generally apply to the detection of CRC in other groups. The model’s performance is superior to current iron deficiency anemia management guidelines, and may help physicians to identify individuals requiring additional clinical evaluation. Conclusions Our model may help to detect CRC earlier in clinical practice.

Identification of patients at risk for colorectal cancer in primary care: an explorative study with routine healthcare data.

Early diagnosis of colorectal cancer (CRC) is likely to reduce burden of disease and improve treatment success. Estimation of the individual patient risk for CRC diagnostic determinants in a primary care setting has not been very successful as yet. The aim of our study is to improve prediction of CRC in patients selected for colonoscopy in the primary healthcare setting using readily available routine healthcare data. A cross-sectional study was carried out in the Julius General Practitioners' Network database. Patients referred for colonoscopy by their general practitioner (GP) between 2007 and 2012 were selected. We evaluated the association between long-term registered patient characteristics, symptoms and conditions, and colonoscopy test results with multivariable logistic regression. Two per cent (2787/140 000) of the patients between 30 and 85 years were found to be newly referred for colonoscopy by their GP, of whom 57 (2%) were diagnosed with CRC. Age 50 years or over, hypertension and the absence of preceding consultations for abdominal pain were independent predictors for CRC and/or high-risk adenomas, with an area under the curve of 0.65. Three factors in routine care data combined might prove valuable in future strategies to improve the prediction of CRC risk in primary care. Improvement in quality and availability of routine care data for research and risk stratification is needed to optimize its usability for prediction purposes in daily practice. Only referring patients at the highest risk for colonoscopy by the GP could decrease superfluous colonoscopies.