Abstract
Objective
To compare the number of contacts to general practice across 11 types of abdominal cancer in the 12 months preceding a diagnosis.
Design
Nationwide register study.
Setting
Danish general practice.
Subjects
Forty-seven thousand eight hundred and ninety-eight patients diagnosed with oesophageal, gastric, colon, rectal, liver, gall bladder/biliary tract, pancreatic, endometrial, ovarian, kidney or bladder cancer in 2014–2018.
Main outcome measures
Monthly contact rates and incidence rate ratios (IRRs) of daytime face-to-face, email and telephone consultations in general practice across different abdominal cancers. The analyses were conducted for each sex and adjusted for age, comorbidity, marital status and education.
Results
Compared to women with colon cancer, women with rectal cancer had the lowest number of contacts to general practice (IRR 12 months pre-diagnostic (IRR–12)=0.86 (95% CI: 0.80–0.92); IRR 1 month pre-diagnostic (IRR–1)=0.85 (95% CI: 0.81–0.89)), whereas women with liver (IRR–12=1.23 (95% CI: 1.09–1.38); IRR–1=1.11 (95% CI: 1.02–1.20)), pancreatic (IRR–12=1.08 (95% CI: 1.01–1.16); IRR1=1.52 (95% CI: 1.45–1.58)) and kidney cancer (IRR–12=1.14 (95% CI: 1.05–1.23); IRR–1=1.18 (95% CI: 1.12–1.24)) had the highest number of contacts. Men showed similar patterns. From seven months pre-diagnostic, an increase in contacts to general practice was seen in bladder cancer patients, particularly women, compared to colon cancer.
Conclusions
Using pre-diagnostic contact rates unveiled that liver, pancreatic, kidney and bladder cancers had a higher and more prolonged use of general practice. This may suggest missed opportunities of diagnosing cancer. Thus, pre-diagnostic contact rates may indicate symptoms and signs for cancer that need further research to ensure early cancer diagnosis.
The majority of cancer patients attend their general practitioner (GP) before diagnosis; however, little is known about the use of general practice across different abdominal cancers.
This study suggests that a potential exists to detect some abdominal cancers at an earlier point in time.
The contact patterns in general practice seem to be shaped by the degree of diagnostic difficulty.
GPs may need additional diagnostic opportunities to identify abdominal cancer in symptomatic patients.
Key points
Introduction
General practitioners (GPs) act as gatekeepers to the rest of the healthcare system in Denmark, except for emergencies, otorhinolaryngologists and ophthalmologists. Thus, most cancer patients attend their GP prior to a diagnosis of cancer [Citation1,Citation2]. The GPs are responsible for referring patients to relevant diagnostic investigations in case of suspicion of cancer. Abdominal symptoms account for approximately 10% of encounters in general practice [Citation3,Citation4], and many symptoms and signs of abdominal cancer mimic each other [Citation4,Citation5]. Therefore, it can be difficult to identify the abdominal cancer causing the presented symptoms [Citation6], and the GP may thus not take appropriate actions, which may lead to missed opportunities for more timely diagnosis of the cancer [Citation7].
Missed opportunities are instances in which post hoc judgement indicates that alternative decisions or actions could have led to a more timely diagnosis, and such opportunities may occur in different phases of the diagnostic process [Citation7]. Thus, delayed diagnosis may be attributable to a missed opportunity [Citation2]. As shorter time to diagnosis is associated with a more favourable prognosis in several types of cancers [Citation8], identifying possibilities to diagnose the cancer at an earlier point in time is an important step to shorten the time to diagnosis, including the time spent in general practice [Citation9,Citation10].
Studies indicate that increased use of healthcare services can be seen as a proxy for presentation of symptoms [Citation11,Citation12], and an increased frequency of healthcare utilisation among cancer patients is seen already several months before their diagnosis compared to patients without cancer [Citation1,Citation13,Citation14]. This increase may represent missed opportunities, indicating opportunities to diagnose some cancers at an earlier point in time. This diagnostic window may vary across different types of abdominal cancers.
It has been suggested that cancers can be categorised into easy-to-diagnose, intermediate-to-diagnose and hard-to-diagnose cancer types according to how difficult the cancer is to suspect and diagnose in general practice; this categorisation is based on the consultation frequency in the year preceding a diagnosis [Citation15]. However, the timing of these consultations is unclear across different abdominal cancers.
We aimed to compare the frequency and timing of contacts to general practice across different abdominal cancer types in the 12 months preceding a diagnosis, to identify possible missed opportunities to diagnose the cancer at an earlier point in time.
Methods
Study design
We conducted a national registry-based cohort study of first-time abdominal cancer patients, who were diagnosed between 1 January 2014 and 31 December 2018. We used the unique Danish civil registration number, which is assigned to all Danish residents at birth or immigration [Citation16], to link the data at an individual level across registers.
Setting
The study was set in Denmark, where a publicly funded healthcare system offers free access to both general practice and hospital care. Almost all Danish residents (>98%) are registered with a specific general practice [Citation17], which they must consult for medical advice.
Study population
Patients eligible for inclusion were first-time cancer patients, aged ≥18 years, registered in the Danish Cancer Registry (DCR) [Citation16] with an abdominal cancer in Denmark in the study period, and with no prior history of cancer (except prior non-melanoma skin cancer). Patients with a valid civil registration number, patients who had lived in Denmark during (at least) the 12 months preceding the date of diagnosis, and patients listed with a general practice were included.
The included abdominal cancer types were defined according to the International Classification of Disease (ICD-10), version 2016, and comprised the following cancer types: oesophageal (C15.0–C15.9), gastric (C16.0–C16.9), colon (C18.0–C18.9), rectal (C20), liver (C22.0–C22.9), gall bladder/biliary tract (C23 and C24.0–C24.9), pancreatic (C25.0–C25.9), endometrial (C54.0–C54.9 and C55), ovarian (C56), kidney (C64 and C65) and bladder cancer (C67.0–C67.9). The date of diagnosis and the ICD-10 diagnosis code were retrieved from the DCR.
Outcomes
The main outcome was number of contacts to general practice. This number included daytime face-to-face consultations (consultations in general practice and home visits), telephone consultations and email consultations. Telephone and email consultations were included because consultations in general practice are increasingly conducted through different communication technologies in Denmark [Citation18]. However, differences in symptom presentation are seen across abdominal cancers [Citation5,Citation19], and these differences may also be reflected in the mode of contact used by abdominal cancer patients to contact general practice. Therefore, a sub-analysis was performed to explore whether differences were observed in the type of contact to general practice across the abdominal cancers.
Data were obtained from the Danish National Health Service Register, which holds information on all types of consultations involving general practice [Citation17]. The data are registered electronically as part of remuneration system.
Covariates
To adjust for differences between groups we included: sex, age, comorbidity, marital status and educational level. Age at the date of diagnosis was modelled through restricted cubic splines with three knots according to Harrell’s recommended percentiles [Citation20]. The Charlson Comorbidity Index (CCI) was calculated based on diagnoses registered in the DNPR in the 10 years preceding study entry (i.e. starting from one year before the date of the cancer diagnosis) and grouped into ‘none’ (CCI score: 0), ‘moderate’ (CCI score: 1–2) and ‘severe’ (CCI score: ≥3). Marital status and educational level were collected from Statistics Denmark [Citation16]. Marital status was categorised into ‘living alone’ and ‘married/cohabitating.’ The highest attained level of education was categorised according to the International Standard Classification of Education (ISCED) into ‘low’ (ISCED level: 1–2), ‘medium’ (ISCED level: 3–4) and ‘high’ (ISCED level: ≥5). Missing information on educational level was seen in 2% of the patients and was recoded as ‘low’ as prior studies have shown that these patients often have lower levels of education [Citation21].
Statistical analysis
Monthly mean contact rates in general practice were calculated for each of the 12 months preceding the date of diagnosis. All patients were included in the analyses and used in the denominator of the rate calculation during all 12 months prior to the diagnosis. The point of increase was defined as the first month prior to diagnosis when the rates were higher than in the month immediately before, and this point was determined by the corresponding confidence intervals (CIs) (i.e. the CIs for two adjacent months did not overlap).
Generalised linear models with log link for the negative binomial family were used to calculate incidence rate ratios (IRRs) to enable comparison of monthly rates of contacts to general practice between different abdominal cancer types. Colon cancer was chosen as the reference cancer because this cancer type has been suggested to belong to the intermediate-to-diagnose group of cancers [Citation15], and is the most common abdominal cancer type in both sexes [Citation22]. Cluster robust variance estimation was applied to account for possible cluster effects at patient level as measurements on the same person were repeated monthly. IRRs were adjusted for age, comorbidity, marital status and educational level. All analyses were performed for each sex separately, as differences in healthcare use have previously been demonstrated to depend on sex [Citation1,Citation23].
The sub-analysis assessed the monthly contact rates and the IRRs for the different abdominal cancersstratified by type of contact (daytime face-to-face consultation, telephone consultation or email consultation) in general practice.
A p value of ≤.05 was considered statistically significant, and estimates were provided with a 95% CI. Analyses were performed using Stata statistical software, version 15 (StataCorp LP, College Station, TX).
Results
A total of 47,898 abdominal cancer patients were included, of which colon cancer patients constituted 31.4%. Women represented 47.0% of the cases; the mean age was 69.5 years (standard deviation (SD)=12.2), 30.5% had comorbidity, and 43.9% had a lower level of education. Men were slightly younger, had more comorbidity, and had higher levels of education than women ().
Table 1. Characteristics of the study population shown by abdominal cancer type and total and stratified by sex (N = 47,898).
Contacts to general practice for different abdominal cancers
A significant increase in contacts to general practice was seen for both sexes from three months prior to the diagnosis across the 11 different abdominal cancer types (); however, the contact rates increased at different relative rates (). For women with e.g. colon cancer, the mean number of contacts increased from 1.21 (95% CI: 1.17–1.25) in the three months before the diagnosis to 2.36 (95% CI: 2.31–2.41) in the last month before the diagnosis (). The results for men demonstrated similar patterns, although men had relatively fewer contacts to general practice than women across all types of abdominal cancers ().
Figure 1. Contacts to general practice across abdominal cancer patients in the 1–12 months preceding the diagnosis. Contacts included daytime face-to-face consultations, telephone consultations and email consultations. Upper part: monthly mean contacts to general practice. Lower part: incidence rate ratios for contacts to general practice compared to colon cancer and adjusted for age, comorbidity, educational level and marital status (95% confidence intervals can be found in Appendix 1). D: date of diagnosis.
Compared to women with colon cancer, women with rectal cancer had fewer contacts to general practice (IRR at 12 months before diagnosis (IRR–12)=0.86 (95% CI: 0.80–0.92); IRR at one month before diagnosis (IRR–1)=0.85 (95% CI: 0.81–0.89)), whereas women with liver cancer (IRR–12=1.23 (95% CI: 1.09–1.38); IRR–1=1.11 (95% CI: 1.02–1.20)), pancreatic cancer (IRR–12=1.08 (95% CI: 1.01–1.16); IRR–1=1.52 (95% CI: 1.45–1.58)) and kidney cancer (IRR–12=1.14 (95% CI: 1.05–1.23); IRR–1=1.18 (95% CI: 1.12–1.24)) had the highest use of general practice ( and Appendix 1, Table S1).
During the last 6–7 months before the diagnosis, an increase in contacts to general practice was also seen in bladder cancer patients compared to colon cancer patients, particularly in women, where the greatest IRR was seen at two months preceding the diagnosis (IRR–2=1.63 (95% CI: 1.52–1.74)) ( and Appendix 1, Table S1).
In both sexes, patients with oesophageal, gastric and gall bladder/biliary tract cancer demonstrated contact rates almost equal to the contact rates for colon cancer (IRR = 1), although an increase in IRR was seen during the last 3–6 months ( and Appendix 1, Table S1). Compared to women with colon cancer, women with endometrial cancer had slightly lower contact rates for several months before the diagnosis, while women with ovarian cancer had higher consultation rates during the last three months preceding the diagnosis ( and Appendix 1, Table S1).
Overall, men showed similar patterns, although the exact estimates were slightly different ( and Appendix 1, Table S1).
Contacts to general practice stratified by type of contact
The sub-analyses displayed similar results as the main analyses; the only difference was that patients with liver cancer had a relatively higher proportion of telephone consultations during all 12 months preceding the diagnosis compared to patients with colon cancer (, and Appendix 1, Table S2).
Figure 2. Contacts to general practice across abdominal cancer patients in the 1–12 months preceding the diagnosis stratified by type of contact (daytime face-to-face consultations, telephone consultation and email consultations). Upper part: monthly mean contacts to general practice. Lower part: incidence rate ratios for contacts to general practice compared to colon cancer and adjusted for age, comorbidity, educational level and marital status. 95% confidence intervals can be found in Appendix 1. D: date of diagnosis.
Discussion
Main findings
In this study of nearly 50,000 first-time abdominal cancer patients, we found a significant increase in contacts to general practice from three months prior to diagnosis. This increase was seen across cancer types and for both sexes. Rectal cancer patients had fewer contacts to general practice than colon cancer patients during all 12 months prior to the diagnosis, whereas patients with liver, pancreatic or kidney cancer had more contacts. A notable increase in contacts to general practice was also found in patients with bladder cancer compared to patients with colon cancer; this increase was particularly seen in women in the last 6–7 months before the diagnosis.
Strengths and limitations
To our knowledge, this is the first study to compare the frequency and timing of contacts to general practice across different abdominal cancer types in the year preceding a diagnosis of cancer. An important strength was the nationwide design and the inclusion of all incident abdominal cancer patients registered in the DCR during a five-year period. An additional strength was the linkage of this data to other Danish National Registers; these registers are known to generate highly valid and complete data [Citation16], which minimised the risk of selection and information bias.
The main limitation of this study was the lack of information on the reasons for contacting general practice, as such data is not available from the registers. Therefore, some of the contacts might have been unrelated to the cancer disease. However, such contacts would have to vary substantially between the different abdominal cancers to explain the findings of this study, which is unlikely, especially as we included age and comorbidity in the adjusted analyses.
Comparison with other studies
Our findings of increased use of general practice prior to an abdominal cancer diagnosis is in line with other studies exploring healthcare use for specific cancer types [Citation1,Citation13,Citation14,Citation23]. However, the timing of the increase in contacts to general practice occurred later in our study than in studies comparing cancer patients with references without cancer. Those studies reported increases in contacts for up to 12 months prior to the diagnosis [Citation1,Citation23]. A likely explanation for this discrepancy is that we compared the contact rates between patients with different cancer types. Hence, the mean contact rate is expected to be elevated at several months prior to diagnosis in our population compared to a population of healthy controls. Nevertheless, our finding of increased number of contacts to general practice in the three months before the diagnosis is in line with another Danish study using a similar methodology [Citation14]. Our comparison of contacts for different cancer types makes it possible to gain new knowledge on how different abdominal cancers are diagnosed. This approach highlights the cancers associated with increased use of general practice for a longer period of time before the diagnosis, which can be seen as a proxy for longer diagnostic intervals.
Variations in the proportion of patients who visited their GP three or more times before referral to hospital have been used to categorise cancer types into groups according to diagnostic difficulty [Citation15,Citation24]. According to this classification, colon cancer and oesophageal cancer is categorised as an intermediate-to-diagnose cancer type, while rectal cancer and endometrial cancer belong to the easy-to-diagnose group [Citation15]. This is in line with our findings that rectal cancer patients had lower mean contact rates than colon cancer patients. Additionally, in our study, patients with pancreatic and kidney cancer had higher mean contact rates than colon cancer patients, which is in line with previous studies that categorised these cancers as a hard-to-diagnose cancer types [Citation15]. In contrast, gastric and ovarian cancers are also categorised as hard-to-diagnose cancers [Citation15], and yet we found similar contact rates for these patients and patients with colon cancer in the present study. Still, our results indicated that abdominal cancers characterised as ‘easy’ are those often presenting with bleeding (e.g. rectal and endometrial cancers) while those characterised as ‘hard’ are the solid abdominal tumours (e.g. kidney cancer). However, many cancers have several possible presentations [Citation5,Citation19], which may challenge the clinical utility of this categorisation. Still, our findings suggest that an opportunity exists to detect some abdominal cancers earlier.
Lyratzopoulos et al. also found a strong effect of sex for bladder cancer, including a doubled increased risk in women with bladder cancer to have more than three consultations before a hospital referral [Citation15]. Our findings support this notable increase in the number of contacts to general practice in female bladder cancer patients in the last 6–7 months before a diagnosis. Despite differences in methodologies, the similar findings in both studies suggest that increased attention should be directed towards women presenting repeatedly in general practice, specifically if they present with urinary tract symptoms.
Clinical implications
Our findings allude to potential missed opportunities across most abdominal cancers. Furthermore, our findings of variations in the use of general practice across the investigated abdominal cancers, suggest that these variations may be caused by different proportions of patients presenting with cancer alarm symptoms across the different abdominal cancer types. This warrants focus on how to improve the diagnostic process and support the GP in the clinical decision-making to improve the chance for more timely diagnosis of cancer. Fast-track cancer patient pathways (CPPs) were implemented in 2007/2008 in Denmark for patients with cancer-alarm symptoms, and followed by a CPP for serious non-specific symptoms and signs in 2012 [Citation6]. However, as most established pathways for cancer diagnosis require specific symptoms or initial GP suspicion of cancer [Citation2,Citation25], a more direct access pathway may be needed for GPs when patients with non-specific or vague symptoms consult [Citation6]. Therefore, novel initiatives aiming to ensure early diagnosis of abdominal cancer may benefit from focussing on development and testing of new referral options for GPs to facilitate the diagnosing of the intermediate- to hard-to-diagnose abdominal cancer types.
Besides better investigation and referral options in general practice, GPs may need clinical decision support tools, such as artificial intelligence (AI) [Citation26] or clearer guidance on effective safety netting for cancer [Citation27]. This could assist the GPs in the initial diagnostic assessment and the diagnostic follow-up. AI could support the GP by recognising and managing cognitive bias to improve the diagnostic accuracy, and safety-netting could help ensure timely and appropriate follow-up when symptoms and signs do not improve [Citation26,Citation27]. Furthermore, point-of-care testing (POCT) is increasingly explored for use in general practice, as POCT may support the early cancer diagnosing in general practice [Citation28–30].
Conclusions
Increased use of general practice was observed for all abdominal cancers in the months before diagnosis, when examining pre-diagnostic contact rates. The time length of the increase varied across different cancer types. Rectal cancer patients had fewer contacts to general practice during all 12 months preceding the diagnosis compared to colon cancer patients, whereas patients with liver, pancreatic or kidney cancer had more contacts during the 12 months. Thus, abdominal cancers that are hard-to-diagnose tended to cause a larger increase in the number of contacts to general practice over many months before the diagnosis. A notable increase was found in women with bladder cancer compared to women with colon cancer. This implies that increased awareness should be addressed to women presenting with urinary tract symptoms in general practice.
Overall, our findings suggest that opportunities may exist to detect some abdominal cancers earlier. These findings call for additional support to assist GPs in diagnosing abdominal cancer. However, approaches are warranted to overcome the methodological challenges in identifying missed opportunities, particularly for cancers which are not associated with clear ‘alarm’ symptoms.
Authorship contributions
NHJ provided the basic idea and contributed to the preparation of the study, performed the statistical analyses, and drafted the manuscript. HJ participated in the design, the data interpretation, and the revision of the manuscript. AZF assisted with the statistical analyses. HGL provided critical input to revisions of the manuscript, and HGR participated in the planning of the study and made crucial revisions to the drafts. PV interpreted the results and provided critical revisions to the manuscript.
All authors read and approved the final version of the manuscript.
Supplemental Material
Download MS Word (50.8 KB)Acknowledgements
We would like to thank data manager Kaare Rud Flarup for helping with the initial data management, Statistics Denmark for facilitating linkage between registers, and Lone Niedziella for linguistic revision.
Ethical considerations: The study was approved and registered in the Record of Processing Activities at the Research Unit for General Practice in Aarhus in accordance with the provisions of the General Data Protection Regulation (GDPR) of the European Union. According to Danish law, the study required no approval from the Committee on Health Research Ethics of the Central Denmark Region as no biomedical intervention was performed. The study was conducted in accordance with the STROBE statement.
Disclosure statement
Henning Gronbaek is a member of the advisory board of the Ipsen Biopharmaceutical Group and has received research grants from Abbvie, Intercept, ARLA, ADS AIPHIA Development Services AG, and the NOVO Nordisk Foundation. The funders had no role in the design or reporting of the study. All other authors declare to have no conflict of interests.
Data availability statement
The datasets included in this study are not publicly available as they contain information that might impair the privacy of the study population. The datasets are stored electronically at Statistics Denmark and can only be accessed by approved collaborators.
Additional information
Funding
References
- Christensen KG, Fenger-Gron M, Flarup KR, et al. Use of general practice, diagnostic investigations and hospital services before and after cancer diagnosis – a population-based nationwide registry study of 127,000 incident adult cancer patients. BMC Health Serv Res. 2012;12:224.
- Jensen H, Torring ML, Olesen F, et al. Cancer suspicion in general practice, urgent referral and time to diagnosis: a population-based GP survey and registry study. BMC Cancer. 2014;14:636.
- Vandvik PO, Kristensen P, Aabakken L, et al. Abdominal complaints in general practice. Scand J Prim Health Care. 2004;22(3):157–162.
- Holtedahl K, Hjertholm P, Borgquist L, et al. Abdominal symptoms and cancer in the abdomen: prospective cohort study in European primary care. Br J Gen Pract. 2018;68(670):e301–e310.
- Koo MM, Swann R, McPhail S, et al. Presenting symptoms of cancer and stage at diagnosis: evidence from a cross-sectional, population-based study. Lancet Oncol. 2020;21(1):73–79.
- Vedsted P, Olesen F. A differentiated approach to referrals from general practice to support early cancer diagnosis – the Danish three-legged strategy. Br J Cancer. 2015;112(Suppl. 1):S65–S69.
- Lyratzopoulos G, Vedsted P, Singh H. Understanding missed opportunities for more timely diagnosis of cancer in symptomatic patients after presentation. Br J Cancer. 2015;112(Suppl. 1):S84–S91.
- Neal RD, Tharmanathan P, France B, et al. Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic review. Br J Cancer. 2015;112(Suppl. 1):S92–S107.
- Weller D, Vedsted P, Rubin G, et al. The Aarhus statement: improving design and reporting of studies on early cancer diagnosis. Br J Cancer. 2012;106(7):1262–1267.
- Rubin G, Berendsen A, Crawford SM, et al. The expanding role of primary care in cancer control. Lancet Oncol. 2015;16(12):1231–1272.
- Ahrensberg JM, Fenger-Grøn M, Vedsted P. Use of primary care during the year before childhood cancer diagnosis: a nationwide population-based matched comparative study. PLOS One. 2013;8(3):e59098.
- Zhou Y, Abel GA, Hamilton W, et al. Imaging activity possibly signalling missed diagnostic opportunities in bladder and kidney cancer: a longitudinal data-linkage study using primary care electronic health records. Cancer Epidemiol. 2020;66:101703.
- Morrell S, Young J, Roder D. The burden of cancer on primary and secondary health care services before and after cancer diagnosis in New South Wales, Australia. BMC Health Serv Res. 2019;19(1):431.
- Jensen H, Vedsted P, Møller H. Consultation frequency in general practice before cancer diagnosis in relation to the patient's usual consultation pattern: a population-based study. Cancer Epidemiol. 2018;55:142–148.
- Lyratzopoulos G, Neal RD, Barbiere JM, et al. Variation in number of general practitioner consultations before hospital referral for cancer: findings from the 2010 National Cancer Patient Experience Survey in England. Lancet Oncol. 2012;13(4):353–365.
- Thygesen LC, Daasnes C, Thaulow I, et al. Introduction to Danish (nationwide) registers on health and social issues: structure, access, legislation, and archiving. Scand J Public Health. 2011;39(7 Suppl.):12–16.
- Andersen JS, Olivarius Nde F, Krasnik A. The Danish National Health Service Register. Scand J Public Health. 2011;39(7 Suppl.):34–37.
- Grønning A, Assing Hvidt E, Nisbeth Brøgger M, et al. How do patients and general practitioners in Denmark perceive the communicative advantages and disadvantages of access via email consultations? A media-theoretical qualitative study. BMJ Open. 2020;10(10):e039442.
- Koo MM, Hamilton W, Walter FM, et al. Symptom signatures and diagnostic timeliness in cancer patients: a review of current evidence. Neoplasia. 2018;20(2):165–174.
- Harrell F. Regression modeling strategies with applications to linear models LR, and survival analysis. 1st ed. New York: Springer Science & Business Media; 2001.
- Denmark S. Uddannelsesstatistikkens manual [Manual of educational statistics]. Copenhagen, Denmark: Statistics Denmark Inc.; 2016.
- Engholm G, Ferlay J, Christensen N, et al. NORDCAN — a Nordic tool for cancer information, planning, quality control and research. Acta Oncol. 2010;49(5):725–736.
- Hansen PL, Hjertholm P, Vedsted P. Increased diagnostic activity in general practice during the year preceding colorectal cancer diagnosis. Int J Cancer. 2015;137(3):615–624.
- Lyratzopoulos G, Wardle J, Rubin G. Rethinking diagnostic delay in cancer: how difficult is the diagnosis? BMJ. 2014;349:g7400.
- Probst HB, Hussain ZB, Andersen O. Cancer patient pathways in Denmark as a joint effort between bureaucrats, health professionals and politicians — a national Danish project. Health Policy. 2012;105(1):65–70.
- Summerton N, Cansdale M. Artificial intelligence and diagnosis in general practice. Br J Gen Pract. 2019;69(684):324–325.
- Evans J, Ziebland S, MacArtney JI, et al. GPs' understanding and practice of safety netting for potential cancer presentations: a qualitative study in primary care. Br J Gen Pract. 2018;68(672):e505–e511.
- CanTest: Investigation of potential for using point of care tests for early cancer diagnosis in primary care; [cited 2021 Apr 8]. Available from: https://www.cantest.org/research-projects/investigation-of-potential-for-using-point-of-care-tests-for-early-cancer-diagnosis-in-primary-care-phd/
- Verbakel JY, Turner PJ, Thompson MJ, et al. Common evidence gaps in point-of-care diagnostic test evaluation: a review of horizon scan reports. BMJ Open. 2017;7(9):e015760.
- Walter FM, Thompson MJ, Wellwood I, et al. Evaluating diagnostic strategies for early detection of cancer: the CanTest framework. BMC Cancer. 2019;19(1):586.