Key message:

Headline figures for the overall proportion of emergency presentations for cancer differ between the two largest studies conducted: 13% and 24%. However, we show that this difference is not necessarily a true one but reflects differences in classifications and methods of counting. Data collected in primary and secondary care have strengths and weaknesses that reflect their source.

Routes to Diagnosis: Comparing multiple studies

New CRUK logo

NCIN Short Report


Improving cancer survival is a key challenge identified in Improving Outcomes: A Strategy for Cancer. Cancer survival estimates in England currently fall below those in many European countries. If cancer survival in England was comparable with the European average then 5,000 or more deaths within five years of diagnosis could be avoided per year.

Identifying and categorising the routes taken by patients to their cancer diagnoses reveals significant survival differences across different presentation routes and helps our understanding of how patients with poor prognosis enter secondary care. This can inform targeted implementation of awareness and early diagnosis initiatives and enable assessment of their success.Multiple studies have examined the proportion of cancers which present by various routes.

The National Audit of Cancer Diagnoses in Primary Care (NACDPC) reported 13% emergency referrals while the Routes to Diagnosis project reported 24% overall emergency presentations (Table 1). The aim of this data briefing is to compare their results in the spirit of a sensitivity analysis which probes the strength of the apparent disagreement between them.

Table 1: Results from the NACDPC and Routes to Diagnosis study with figures aggregated into broader categories (figures may not sum to 100% due to rounding). Detailed definitions of the referral types can be found in the NACDPC report (Rubin et al, 2011) and Routes to Diagnosis paper (Elliss-Brookes et al, 2012).

RvR table 1

*NACDPC figures rescaled to add 5% screening cases.


The two largest studies, Routes to Diagnosis and the NACDPC, are compared directly, alongside a selection of smaller studies. The studies vary in many respects, including their study populations, study periods, case finding and categorisation of possible routes to diagnosis. We compare Routes to Diagnosis and the NACDPC by aggregating reported categories, as broadly as possible, to: emergency routes; non-emergency and non-private routes that are initiated by the GP; routes that do not involve NHS secondary care; and routes via the screening service. The NACDPC excluded screen detected cases (where these could be identified) so to enable a direct comparison these have been assumed to equal 5% and other proportions have been rescaled (for the figures for all cancers combined across males and females).

The choice of aggregation showcased might be seen as unsophisticated, with all routes aggregated to the broader category in which they most plausibly belong. It is however based on the reasonable assumption that missing or uncertain data in primary care indicates a secondary care route, and vice versa. For instance, it is judged that data about referrals from practices are well known to practices and hence missing referral data in the NACDPC indicates a non-practice and likely emergency presentation. Conversely, it is judged that emergency admissions are well known to secondary care and can be firmly linked to resulting diagnoses meaning that missing or uncertain data in Routes to Diagnosis is likely to indicate a GP referral as the start of the patient pathway.

Table 2: Cross tabulation showing the referral type by place of presentation, National Audit of Cancer Diagnosis in Primary Care.

RvR table 2


The proportions of referrals reported in both the NACDPC and in the Routes to Diagnosis study are compared in Table 1. There is a good agreement in the proportions for the broadly aggregated routes. A breakdown of the referral type by the place of presentation from the NACDPC is shown in Table 2.

Ten per cent of patients (1,964) first presented at the GP surgery and were subsequently referred as an emergency - comparable to the figure from Routes to Diagnosis of seven per cent (NCIN 2013). Five per cent of patients (850) presented at A&E with the majority of these being classified as an "Emergency", "Not referred by practice" or "Not known" referral type. Table 3 shows a wider comparison to other studies.

The agreement between Routes to Diagnosis and the NACDPC seen across all cancers also holds reasonably well for individual tumour types, and both appear to agree well with other published studies.Comparison of the Two Week Wait (TWW) and non-TWW GP referrals (as originally published) shows the Routes to Diagnosis figures at the lower end and the NACDPC at the upper end of those given by other studies.

Table 3: Proportion of presentations by various routes by tumour type in the Routes to Diagnosis study, the NACDPC and other studies in the literature. "TWW" refers to Two Week Wait referrals, and All GP Referrals and All Emergency columns are defined as in Table 1, except that no rescaling for screening has been applied.

RvR table 3


Pathways taken prior to a cancer diagnosis can be complex - patients may move between primary and secondary care more than once on their route to diagnosis. This complexity is demonstrated by the variation of referral route with place of presentation in the NACDPC and shown in Table 2. Representing these complex pathways as summary figures demands that choices are made in operational definitions that may differ from study to study. In particular the categorisation of "Emergency Presentation" in Routes to Diagnosis is likely to be more inclusive than the "Emergency" referrals that the NACDPC uses. Also, for all studies there are multiple mechanisms by which under-ascertainment may occur, which differ between primary and secondary care. These could lead both to differences in case finding and in the completeness with which different data fields are collected. We should therefore accept that some degree of difference between different studies is to be expected.

The approximate agreement, 10% vs 7%, in the overall proportion of emergency referrals from GPs (i.e., excluding self-referrals via A&E) and the fact that these routes represent a minority of cases in Routes to Diagnosis but a majority in the NACDPC is suggestive: we can interpret this as evidence that many cases in the NACDPC are emergency presentations via A&E, as this would then make the two studies consistent.

The agreement in scale between private referrals in the NACDPC and the unknowns in Routes to Diagnosis is interesting and suggests that a large proportion of Unknown presentation routes in Routes to Diagnosis are in fact private referrals. This would explain the fact that 'Unknown' routes have 1-year survival is roughly equivalent to persons following known care pathways. It is also plausible that more affluent people, and those of working age, preferentially choose private referrals. This would also explain the socio-demographic variation in unknown routes observed in the Routes to Diagnosis study (NCIN 2012).

The nature of Inpatient Elective and Other Outpatient routes in the Routes to Diagnosis study was not fully explained. Broadly classifying them as a GP referral implies that a GP referral was the action that ultimately started the route to diagnosis for these cases, though potentially via a complex pathway. The aggregation we propose is simple but sensible and allows the results of the Routes to Diagnosis study to be seen as complementary to studies in primary care, as noted by Rubin et al, 2013. However, the aggregation of different operational definitions used here does not allow for incidental diagnoses or for referrals via (for example) walk-in centres. This indicates that further, more detailed, exploration of the subject is desirable.

In conclusion, what appears to be a difference between the NACDPC and Routes to Diagnosis is not necessarily a true one but one that reflects differences in classifications and methods of counting. Data collected in primary and secondary care have strengths and weaknesses that reflect their source. Great progress in further understanding the referral and diagnosis pathways would come from directly comparing primary and secondary care data at a patient level.


We thank Kathy Elliott, Professor Greg Rubin and Georgios Lyratzopoulos for permission to include data from the NACDPC and/or comments on the report.


Barrett J and Hamilton W (2005) Pathways to the diagnosis of prostate cancer in a British city. A population-based study. Scand J Urol Nephrol, 39: 267-270

Barrett J, Jiwa M, Rose P and Hamilton W (2006) Pathways to the diagnosis of colorectal cancer: an observational study in three UK cities BMC Family Practice 23: 15-19doi:10.1093/fampra/cmi093

Barrett J and Hamilton W (2008) Pathways to the diagnosis of lung cancer in the UK: a cohort study. BMC Family Practice 9:31 doi:10.1186/1471-2296-9-31

Blick C, Bailey D, Haldar N, Bdesha A, Kelleher J, Muneer A (2010) The impact of the two-week wait rule on the diagnosis and management of bladder cancer in a single UK institution. Ann R Coll Surg Engl 92: 46–50 doi 10.1308/003588410X12518836440207

Elliss-Brookes L, McPhail S, Ives A, Greenslade M, Shelton J, Hiom S, Richards M (2012) Routes to Diagnosis for cancer - Determining the patient journey using multiple routine datasets. BJC 107: 1220-1226

Rubin G, Neal R, Abel G, and Lyratzopoulos G (2013) Integrated research efforts are needed to better understand how to reduce the proportion of patients with cancer who are diagnosed as emergencies. BJC, 108(7):1550-1NCIN (2012) Routes to diagnosis 2006-2008 NCIN Information supplement

NCIN (2013) Routes to Diagnosis: Exploring Emergency Presentations.

Neal R, Allgar V, Nasreen A, Leese B, Heywood P, Proctor G, Evans J (2007) Stage, survival and delays in lung, colorectal, prostate and ovarian cancer: comparison between diagnostic routes. British Journal of General Practice 57: 212–219

Rubin G, McPhail S, Elliott K (2011) National Audit of Cancer Diagnosis in Primary Care. Royal College of General Practitioners: London, England

Thorne K, Hutchings H, Elwyn G (2009) The Two-Week Rule for NHS Gastrointestinal Cancer Referrals: A Systematic Review of Diagnostic Effectiveness. The Open Colorectal Cancer Journal, 2: 27-33


The NCIN developed the Routes to Diagnosis methodology:Routes to diagnosis

Cancer Research UK CancerStats

The National Cancer Intelligence Network (NCIN) is a UK-wide partnership operated by Public Health England. The NCIN coordinates and develops analysis and intelligence to drive improvements in prevention, standards of cancer care and clinical outcomes for cancer patients.