Abstract
Background.
We performed a systematic review of diagnostic studies of symptomatic patients in primary care to quantify the risk of brain/central nervous system (CNS) cancer in patients presenting in primary care with symptoms that may indicate brain/CNS cancer.
Objective.
To quantify the risk of brain/CNS cancer in symptomatic patients presenting in primary care.
Methods.
We searched Medline, Premedline, Embase, the Cochrane Library, Web of Science and ISI Proceedings (1980 to August 2014) and PsychInfo (1980 to February 2013) for diagnostic studies of symptomatic adult patients in primary care. Study quality was assessed using QUADAS-II and data were extracted to calculate the positive predictive values (PPVs) of symptoms, singly or in combination, for brain/CNS cancer.
Results.
Six studies with 159938 patients were included. The PPVs of single symptoms were very low with only ‘new-onset seizure’ being above 1% in patients aged 18 years and above, rising to 2.3% in patients aged 60–69 years. In patients aged 15–24 years, the PPVs for the individual symptoms were also very low, with the highest, also for seizure, being 0.024%, similar to that in children aged 0–14 years of 0.02%. For symptom combinations, none of the PPVs were above 0.39%.
Conclusions.
All the symptoms of brain tumours are individually low risk, apart from new-onset epilepsy. This provides a real diagnostic problem, as brain tumours have all the expected features seen with cancer diagnostic delay, with high proportions presenting as an emergency and having had multiple primary care consultations before referral, and the prognosis is poor. Improving these metrics can only be done by liberalizing investigation, although the health economics of that strategy is undetermined.
Keywords: Brain neoplasms, central nervous system neoplasms, diagnosis, primary health care, symptoms.
Introduction
Brain tumours account for ~2% of all new tumours in the UK, with ~4500 new diagnoses each year ( 1 ). The overall annual incidence is 7 per 100000. Primary brain tumours represent ~60% of these and are usually gliomas. Meningiomas represent a further quarter: they are usually benign, though can produce considerable morbidity, depending on the precise location. Endocrine tumours, especially arising in the pituitary, equate to ~11% of the total ( 2 ). Secondary brain tumours are approximately three times more common than primary ones ( 3 ); in most such patients, the primary cancer has already been diagnosed, though a small proportion first present with cerebral metastases ( 4 ). Malignancies of central nervous tissue outside the brain may also occur, though these are much rarer, and have received less research attention. The estimated cost of brain tumours in Europe in 2010 was €5.2 billion ( 5 ).
Several symptoms of primary brain tumours have been described, usually arising from increased intracranial pressure or damage to cerebral tissue. Up to 70% of patients have a headache during the course of their illness, particularly in the final stages of their disease ( 6 ). The incidence of headache at the time of diagnosis is between 23% and 56% ( 6–8 ). Epileptic seizures may also occur, especially in younger patients ( 9 , 10 ). Other reported symptoms of brain tumours include confusion, dysphasia, hemiplegia, hemianopia, motor weakness, personality change and memory loss ( 11 ).
Most patients with symptoms of brain tumours present to primary care, although patient with new-onset epilepsy may present directly to emergency departments. As most symptoms associated with brain tumours are also common in patients without tumours, the decision whether—or not—to investigate for a possible tumour is difficult. This difficulty is compounded by the main investigation being imaging, which is relatively expensive; furthermore, imaging may identify innocent lesions, the so-called ‘incidentalomas’. Imaging is generally performed as a specialist investigation, though schemes extending this to primary care clinicians have been implemented ( 12 , 13 ). We sought to quantify the risk of brain cancer in patients presenting in primary care with symptoms that may indicate brain/central nervous system (CNS) cancer, in a systematic review as part of a revision of UK cancer guidance ( 14 ).
Methods
Criteria for considering studies for this review
We included only studies that were performed on a series of unselected or randomly selected patients presenting to primary care with symptoms that might indicate brain/CNS cancer. The studies could be prospective or retrospective, but had to report follow-up data for the included patients indicating whether they had brain/CNS cancer or whether their symptom was of a benign origin. The follow-up data could be obtained from tests or medical records. Diagnostic case–control studies where cases were patients with brain/CNS cancer and controls were (matched) patients without brain/CNS cancer that reported the prevalence of the symptoms in both patient groups and the outcome of interest [positive predictive value (PPV); see also below] were also included. We did not include studies reporting only on high-risk patients or studies conducted in settings other than primary care, such as secondary or tertiary care, or studies reporting only on patients with brain/CNS cancer.
Search methods for identification of studies
We searched Medline, Premedline, Embase, the Cochrane Library, Web of Science (SCI and SSCI) and ISI Proceedings from 1980 to 12 August 2014 as well as PsychInfo (1980 to 20 February 2013) using the search strategy outlined in the online Supplementary Material . The search was performed by one of the authors (SB) who also screened the initial search results, excluding all obviously irrelevant studies. A second author (MSH) then screened the titles and abstracts of the remaining records, excluding further irrelevant studies and examining the full text of all potentially relevant studies. The final lists of included and excluded studies were agreed in consensus between two of the authors (MSH and WH).
Data collection and analysis
Data extraction and quality assessment of the included studies was performed by one author (MSH). For each included study, the following characteristics were extracted: study design, inclusion/exclusion criteria, setting, patient characteristics (number, age, gender, country, any other relevant characteristics reported such as relevant history or comorbidities), definition of symptom, method of verification of diagnosis and any other relevant details reported in the studies. The risk of different biases associated with the included studies was assessed using the QUADAS-II tool for each of the included studies ( 15 ). For each reported symptom, we extracted the PPV or, where this was not reported, the number of patients with the symptoms who had brain/CNS cancer [true positives (TP)] and the number of patients with the symptoms who did not have brain/CNS cancer [false positives (FP)]. From these data, we calculated PPVs that formed the basis of the risk estimate using the conventional formula of TP/TP + FP. If three or more studies reported a given symptom, we aimed to meta-analyse the results to provide a summary estimate indicating the risk of brain/CNS cancer associated with each symptom. However, meta-analysis was not feasible due to the low number of studies reporting on individual symptoms and the variation in their study design.
Results
Results of the search
The search of all the databases identified 7555 (before de-duplication) possibly relevant papers of which 7525 papers were excluded based on title/abstract: 30 papers were obtained for full text review. Six of these 30 papers were included in this review ( 16–21 ), while 24 papers were excluded for the following reasons: follow-up period not deemed sufficient ( N = 2) ( 22 , 23 ); narrative review ( N = 6); no or insufficient verification of final diagnosis ( N = 4); not patients presenting to primary care ( N = 3); setting not primary care ( N = 1); PPVs not reported or calculable ( N = 3); case report ( N = 3) and did not include any new data ( N = 2).
Characteristics and methodological quality of included studies
Table 1 provides a summary of the characteristics of the included studies and the online Supplementary Materials provide a detailed description and assessment of each study. The studies included a total of 159938 patients with all 6 conducted in the UK. The reference standards employed in the studies were all follow-up. The studies reported on a number of single symptoms and symptom combinations, in children ( 17 , 19 , 21 ), teenagers and young adults ( 18 ) or adult ( 16 , 20 ) populations.
Table 1.
Characteristics of the included studies on patients presenting to primary care with symptoms that might indicate brain/CNS cancer including method of verification of malignant/benign diagnosis
| Study | Country | Design | Total number eligible | Number of patients with brain/CNS cancer | Reference standard |
|---|---|---|---|---|---|
| Ansell et al . (2010) (17) | UK | National population-based case–control study of children aged 1–14 years | 480 | 195 (mean age of patients = 7.31 years; SE = 0.27 years) | Diagnosis of brain tumour in their primary care record |
| Dommett et al . (2013) (18,19) | UK | Population-based nested case– control study using data from the GPRD from children and young adults aged 0–24 years | 5634 | 424 (0–4 years: N = 73; 5–14 years: N = 197; 15–24 years: N = 154) | Diagnosis of brain/CNS cancer in the GPRD |
| Hamilton and Kernick (2007) (20) | UK | Population-based nested case– control study using data from the GPRD from adults aged ≥18 years | 19570 | 2397 (of a total of 3505 cases with brain tumours, the remainder were benign; 584/2397 malignant tumours occurred in patients aged 60–69 years) | Diagnosis of brain tumour in the GPRD |
| Kernick et al . (2008) (16) | UK | Retrospective, consecutive patients using data from the GPRD from adults aged ≥18 years | 85679 | 107 (<50 years: N = 37; ≥50 years: N = 70) | 1-year follow-up |
| Kernick et al . (2009) (21) | UK | Retrospective, consecutive patients using data from the GPRD from children aged 5–17 years | 48575 | 13 | 1-year follow-up |
GPRD, UK General Practice Research Database (now renamed the Clinical Practice Research Datalink); SE, standard error.
Table 2 summarizes the risk-of-bias and applicability assessments for each of the included studies. The main bias and applicability concerns relating to patient selection were that this was not clearly consecutive or random in four of the studies ( 17–20 ), with one of the studies including non-malignant brain tumours ( 20 ).
Table 2.
Risk-of-bias assessment for the included trials
| Risk of bias | Applicability concerns | ||||||
|---|---|---|---|---|---|---|---|
| Study | Patient selection | Index test | Reference standard | Flow and timing | Patient selection | Index test | Reference standard |
| Ansell et al . (2010) (17) | − | + | + | + | + | + | + |
| Dommett et al . (2013) (18,19) | − | + | + | + | + | + | + |
| Hamilton and Kernick (2007) (20) | − | + | + | + | ? | + | + |
| Kernick et al . (2008) (16) | + | + | + | + | + | + | + |
| Kernick et al . (2009) (21) | + | + | + | + | + | + | + |
‘+’ denotes low risk of bias/concerns about applicability; ‘?’ denotes unclear risk of bias/concerns about applicability; ‘−’ denotes high risk of bias/concerns about applicability.
Findings
Single symptoms
Tables 3–5 list the PPVs for single symptoms in adult, teenage and young adult, and child patients. The PPVs of single symptoms were generally very low with only ‘new-onset seizure’ being above 1% in patients aged 18 years and above, rising to 2.3% in patients aged 60–69 years. Table 4 also shows that in patients aged 15–24 years the PPVs for the individual symptoms were very low, with the highest, also for seizure, being 0.024%, similar to that for children aged 0–14 years of 0.02%.
Table 3.
PPVs for brain/CNS cancer for single symptoms in adult patients in descending order of frequency among the patients with brain/CNS cancer
| Study | Symptom(s) | Patient group, age in years | Frequency of symptom in patients with brain/CNS cancer | PPV% (95% CI) |
|---|---|---|---|---|
| Hamilton and Kernick (2007) (20) | Headache | Patients ≥ 18 ( N = 623) | 362/3505 = 10% | 0.09 (0.08–0.1) |
| Patients 60–69 ( N = 54) | 54/767 = 7.0% a | 0.12 (0.09–1.7) a | ||
| Kernick et al . (2008) (16) | Undifferentiated headache | Patients ≥ 18 ( N = 63921) | Not applicable, but PPV based on 97/63921 b | 0.15 (0.12–0.19) |
| Patients < 50 ( N = 40866) | Not applicable, but PPV based on 32/40866 b | 0.08 (0.05–0.11) | ||
| Patients ≥ 50 ( N = 23055) | PPV based on 65/23055 b | 0.28 (0.22–0.36) | ||
| Kernick et al . (2008) (16) | Primary headache | Patients ≥ 18 ( N = 21758) | PPV based on 10/21758 b | 0.045 (0.023–0.088) |
| Patients < 50 ( N = 16282) | PPV based on 5/16282 b | 0.03 (0.01–0.08) | ||
| Patients ≥ 50 ( N = 5476) | PPV based on 5/5476 b | 0.09 (0.03–0.23) | ||
| Hamilton and Kernick (2007) (20) | Motor loss | Patients ≥ 18 ( N = 1039) | 308/3505 = 8.8% | 0.026 (0.024–0.03) |
| Hamilton and Kernick (2007) (20) | New-onset seizure | Patients ≥ 18 ( N = 162) | 154/3505 = 4.4% | 1.2 (1–1.4) |
| Patients 60–69 ( N = 35) | 35/808 = 4.3% a | 2.3 (1.6–3.3) a | ||
| Hamilton and Kernick (2007) (20) | Confusion | Patients ≥ 18 ( N = 156) | 109/3505 = 3.1% | 0.2 (0.16–0.24) |
| Hamilton and Kernick (2007) (20) | Weakness | Patients ≥ 18 ( N = 137) | 95/3505 = 2.7% | 0.14 (0.11–0.18) |
| Hamilton and Kernick (2007) (20) | Memory loss | Patients ≥ 18 ( N = 101) | 37/3505 = 1.1% | 0.036 (0.026–0.052) |
| Hamilton and Kernick (2007) (20) | Visual disorder | Patients ≥ 18 ( N = 97) | 35/3505 = 1% | 0.035 (0.025–0.051) |
a Not reported in the paper but information available from authors.
b Strictly, the PPV is not applicable, but it has been estimated from the numbers given.
Table 5.
PPVs for brain/CNS cancer for single symptoms in children in descending order of frequency among the patients with brain/CNS cancer
| Study | Symptom(s) | Patient group, age in years | Frequency of symptom in patients with brain/CNS cancer | PPV % (95% CI) |
|---|---|---|---|---|
| Dommett et al . (2013) (19) | Headache 0–3 months before diagnosis | All CNS cancer patients and controls aged 0–14 | 58/270 = 21.5% | 0.03 (0.02–0.06) |
| Kernick et al . (2009) (21) | Headache (any type) | All included patients aged 5–17 | PPV based on 13/48575 a | 0.03 (0.01–0.05) |
| Kernick et al . (2009) (21) | Primary headache | PPV based on 0/9321 a | 0 (0–0.05) | |
| Kernick et al . (2009) (21) | Undifferentiated headache | PPV based on 13/38705 a | 0.03 (0.02–0.06) | |
| Dommett et al . (2013) (19) | Vomiting 0–3 months before diagnosis | All CNS cancer patients and controls aged 0–14 | 55/270 = 20.4% | 0.04 (0.02–0.07) |
| Dommett et al . (2013) (19) | Abnormal movement 0–3 months before diagnosis | 32/270 = 11.9% | 0.11 (0.03–0.35) | |
| Dommett et al . (2013) (19) | Visual symptoms 0–3 months before diagnosis | 21/270 = 7.8% | 0.07 (0.02–0.24) | |
| Dommett et al . (2013) (19) | Pain 0–3 months before diagnosis | 14/270 = 5.2% | 0.03 (0.01–0.08) | |
| Dommett et al . (2013) (19) | Seizure 0–3 months before diagnosis | 10/270 = 3.7% | 0.02 (0.01–0.06) | |
| Dommett et al . (2013) (19) | ≥3 consultations | 137/270 = 50.7% | 0.01 (0–0.01) |
a Strictly, the PPV is not applicable but it has been estimated from the numbers given.
Table 4.
PPVs for brain/CNS cancer for single symptoms in teenage–young adult patients in descending order of frequency among the patients with brain/CNS cancer
| Study | Symptom(s) | Patient group, age in years | Frequency of symptom in patients with brain/CNS cancer | PPV% (95% CI) |
|---|---|---|---|---|
| Dommett et al . (2013) (18) | Headache | All included cases and controls aged 15–24 | 33/154 = 21.4% | 0.015 (0.0077–0.0276) |
| Seizure | 18/154 = 11.7% | 0.024 (0.0082–0.0695) | ||
| Visual symptoms | 13/154 = 8.4% | 0.07 a | ||
| Vomiting | 11/154 = 7.1% | 0.012 (0.0041–0.031) | ||
| Pain | 11/154 = 7.1% | 0.0029 (0.0014–0.006) | ||
| ≥3 consultations | 73/154 = 47.4% | 0.0023 (0.0019–0.0029) |
a The authors did not report a PPV as no controls had this feature. For this review, one control had been re-assigned to having visual symptoms, allowing a broad estimation of the PPV.
Symptom pairs
In addition to single symptoms, two of the studies also reported information on PPVs for symptom combinations: Hamilton and Kernick ( 20 ) found that headache in combination with any other of the symptoms examined in their study was associated with a PPV of 0.39% [95% confidence interval (CI): 0.31–0.48] in patients aged 18 years or above. In children aged 0–14 years, Ansell et al . ( 17 ) reported that vomiting in combination with unsteadiness or visual difficulties had PPVs of 0.15% (95% CI: 0.01–0.1) and 0.088% (95% CI: 0.005–0.6), respectively, and that the PPV of headache in combination with unsteadiness was 0.085% (95% CI: 0.005–0.6). All other symptom combinations examined by Ansell et al . ( 17 ) had even lower PPVs.
Conclusions
To our knowledge, this is the first systematic review of the clinical features of brain/CNS cancer in primary care. Many of the features reported from secondary care studies were also relevant in primary care, though the likelihood of cancer with particular features was generally lower in primary care, reflecting the lower incidence of cancer in that population compared to specialist practice. In adults and teenagers, seizures were the symptom with the highest risk, but these risks were still small, with a highest estimate in adults of 2.3% in patients aged 60–69, and they were very small in other age groups, being below 1 in 1000 for children, teenagers or young adults. More common features, such as headache, all represented much lower risks of brain/CNS cancer.
Strengths and weaknesses of the review
This review followed best practice methods ( 24 ). In particular, the setting for the inclusion of studies was primary care. This was crucial for the clinical question to be answered—which patients presenting to primary care may have a brain/CNS cancer, and so may benefit from imaging? Reviews including patients in the referred population generally find stronger associations between symptoms and disease, and are much less helpful in informing investigation or referral decisions.
As with any review, the findings depend upon the quality of the original studies. The more recent ones using electronic research databases were high quality, and of large size, despite the relative rarity of brain/CNS cancer. Several used case–control methods, which can lead to bias from patient selection. In this case, however, all patients present in the GPRD were used, reducing this concern. Two small studies conducted in US-based emergency departments were excluded due to insufficient follow-up periods (1–4 weeks and 1–7 months, respectively) ( 22 , 23 ). Emergency departments are generally considered to be primary care, though the spectrum of disease seen in them differs from usual general or family practice. It was debatable whether these two studies should have been included in this review. However, as no cancers were found in either study, their results are relatively uninformative, except to demonstrate that both weakness and dizziness appear to be very low risk symptoms.
Implications of the findings for clinical practice
The commonest symptom of brain tumours in primary care is headache. This was recorded in the primary care records of 10% of adults in the year before diagnosis of a brain tumour, and of over 20% of children and teenagers and young adults. It is probable that these represent under-estimates, as not every symptom is recorded in medical records, and some are recorded in a part of the notes that is not accessible routinely by researchers ( 25 ). This under-recording is likely to have affected controls as well, so the PPVs (which are based on the ratio between symptom frequency in cases and symptom frequency in controls) are likely to be reasonably accurate. In all age groups, PPVs for headache were very low – and lower still if the doctor was able to make a diagnosis of primary headache ( 21 ). Even when the headache was accompanied by a second symptom, the risks of a brain tumour remained below 1% for both adults and children ( 17 , 20 ). This should help support clinicians discuss the merits and demerits of possible imaging, particularly when a patient with a headache specifically requests it.
In contrast, new-onset epileptic seizures were both less common than headache and of higher risk. Fewer than 5% of adults with a brain tumour had a preceding seizure—and under-recording of this symptom is much less likely than under-recording of headache ( 20 ). Young children had a similar low prevalence of seizures ( 19 ), whereas teenagers and young adults had the highest, at 11.7% ( 18 ). These figures for prevalence of epilepsy with brain tumours are similar to a population study using routine Swedish pathology data ( 10 ). As epileptic seizures are common in patients without brain tumours, PPVs were relatively low, with the highest value being 2.3% in the 60–69 age group. Current UK guidance for a new-onset seizures suggests MRI imaging, except where idiopathic generalized or benign focal epilepsy have been diagnosed ( 26 ). Such imaging will identify a brain tumour, as well as other rarer structural causes of epilepsy.
Brain tumours may also present with a progressive loss of neurological function, which may have a motor component, a sensory component or loss of higher cerebral functions, such as memory. Structural imaging is recommended in suspected dementia, in part to identify other pathologies, as well as differentiation into dementia subtypes ( 27 ). This review identified very low PPVs for brain tumours with motor symptoms, memory loss or visual disorders—in all age groups. Even so, it can be argued that imaging is of value, not just for identification of possible malignancy, but for other possible benign conditions causing the symptoms. Historically in the UK, cerebral imaging has been the province of specialists, though schemes of primary care access to scanning have been established. The yield of treatable lesions has been low, commensurate with the findings of this review ( 13 , 28 ). The health economics of primary care imaging compared to specialist imaging have not been reported, but the former is likely to be cheaper.
Two aspects of this review suggest possible future research topics. Although the PPVs of headache plus a second symptom were low, it remains possible some headache combinations are truly high risk. This is important because headache is the most prevalent symptom in brain tumours, yet the PPVs do not support investigation for headache per se . Secondly, the benefits of expediting brain tumour diagnosis are largely unknown. Even if mortality gains from expedited diagnosis are small (and this is not certain), morbidity benefits may be considerable.
This review demonstrated that all symptoms of brain tumours are individually low risk, with the exception of new-onset epilepsy. This provides a real diagnostic problem, as brain tumours have all the expected features seen with cancer diagnostic delay. A high proportion present as an emergency ( 29 ); a high proportion has multiple primary care consultations before referral ( 30 ) and the prognosis is poor. Improving these metrics can only be done by liberalizing investigation, though as yet there is not the health-economic evidence to support doing so.
Supplementary material
Supplementary material is available at Family Practice online.
Declaration
Funding: departmental resources.
Ethical approval: none.
Conflict of interest: William Hamilton is an associate editor of Family Practice . MSH, WH and SB have no additional conflicts of interest to declare.
Supplementary Material
Acknowledgements
This article presents a systematic review undertaken as part of the 2015 NICE guideline on ‘Suspected cancer: Recognition and referral’, which was developed by the National Collaborating Centre for Cancer (NCC-C). The NCC-C receives funding from the National Institute for Health and Care Excellence (NICE). The authors would like to thank the National Institute for Health and Care Excellence (NICE), the National Collaborating Centre for Cancer and the Guideline Development Group for the ‘Suspected cancer: Recognition and referral’ NICE Guideline (2015). The views expressed in this publication are those of the authors and not necessarily those of the NHS, NICE, the NIHR or the Department of Health in England.
References
- 1. McKinney PA . Brain tumours: incidence, survival, and aetiology . J Neurol Neurosurg Psychiatry 2004. ; 75 ( suppl 2 ): ii12 – 7 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Brown K . Presenting brain tumour statistics to the public: unravelling complexity . In: NCIN Cancer Outcomes Conference, BRIGHTON, UK, 2013 . National Cancer Intelligence Network, London, UK .
- 3. Davis FG, Dolecek TA, McCarthy BJ, Villano JL . Toward determining the lifetime occurrence of metastatic brain tumors estimated from 2007 United States cancer incidence data . Neuro-Oncology 2012. ; 14 : 1171 – 7 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Giordana MT, Cordera S, Boghi A . Cerebral metastases as first symptom of cancer: a clinico-pathologic study . J Neuro-Oncol 2000. ; 50 : 265 – 73 . [DOI] [PubMed] [Google Scholar]
- 5. Gustavsson A, Svensson M, Jacobi F, et al. . Cost of disorders of the brain in Europe 2010 . Eur Neuropsychopharmacol 2011. ; 21 : 718 – 79 . [DOI] [PubMed] [Google Scholar]
- 6. Suwanwela N, Phanthumchinda K, Kaoropthum S . Headache in brain tumor: a cross-sectional study . Headache 1994. ; 34 : 435 – 8 . [DOI] [PubMed] [Google Scholar]
- 7. Jaeckle K . Causes and management of headaches in cancer patients . Oncology 1993. ; 7 : 27 – 31 . [PubMed] [Google Scholar]
- 8. Snyder H, Robinson K, Shah D, Brennan R, Handrigan M . Signs and symptoms of patients with brain tumors presenting to the emergency department . J Emerg Med 1993. ; 11 : 253 – 8 . [DOI] [PubMed] [Google Scholar]
- 9. Lowry JK, Snyder JJ, Lowry PW . Brain tumors in the elderly: recent trends in a Minnesota cohort study . Arch Neurol 1998. ; 55 : 922 – 8 . [DOI] [PubMed] [Google Scholar]
- 10. Schwartzbaum J, Jonsson F, Ahlbom A, et al. . Prior hospitalization for epilepsy, diabetes, and stroke and subsequent glioma and meningioma risk . Cancer Epidemiol Biomark Prev 2005. ; 14 : 643 – 50 . [DOI] [PubMed] [Google Scholar]
- 11. Kaba SE, Kyritsis AP . Recognition and management of gliomas . Drugs 1997. ; 53 : 235 – 44 . [DOI] [PubMed] [Google Scholar]
- 12. Department of Health . Direct Access to Diagnostic Tests for Cancer. Best Practice Referral Pathways for General Practitioners . 2012. . http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_133510 (accessed on 15 April 2015).
- 13. Simpson GC, Forbes K, Teasdale E, Tyagi A, Santosh C . Impact of GP direct-access computerised tomography for the investigation of chronic daily headache . Br J Gen Pract 2010. ; 60 : 897 – 901 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. NICE . Suspected Cancer: Recognition and Referral (NG12) . London, UK: : NICE; , 2015. . [Google Scholar]
- 15. Whiting PF, Rutjes AWS, Westwood ME, et al. . QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies . Ann Intern Med 2011. ; 155 : 529 – 36 . [DOI] [PubMed] [Google Scholar]
- 16. Kernick D, Stapley S, Goadsby PJ, Hamilton W . What happens to new-onset headache presented to primary care? A case-cohort study using electronic primary care records . Cephalalgia 2008. ; 28 : 1188 – 95 . [DOI] [PubMed] [Google Scholar]
- 17. Ansell P, Johnston T, Simpson J, et al. Brain tumor signs and symptoms: analysis of primary health care records from the UKCCS . Pediatrics 2010. ; 125 : 112 – 9 . [DOI] [PubMed] [Google Scholar]
- 18. Dommett RM, Redaniel MT, Stevens MC, Hamilton W, Martin RM . Features of cancer in teenagers and young adults in primary care: a population-based nested case-control study . Br J Cancer 2013. ; 108 : 2329 – 33 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Dommett RM, Redaniel T, Martin RM, Stevens MCG, Hamilton W . Risk of childhood cancer with symptoms in primary care: a population-based case-control study . Br J Gen Pract 2013. ; 63 : e22 – 9 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Hamilton W, Kernick D . Clinical features of primary brain tumours: a case-control study using electronic primary care records . Br J Gen Pract 2007. ; 57 : 695 – 9 . [PMC free article] [PubMed] [Google Scholar]
- 21. Kernick D, Stapley S, Campbell J, Hamilton W . What happens to new onset headache in children that present to primary care? A case-cohort study using electronic primary care records . Cephalalgia 2009. ; 28 : 1188 – 95 . [DOI] [PubMed] [Google Scholar]
- 22. Herr RD, Zun L, Mathews JJ . A directed approach to the dizzy patient . Ann Emerg Med 1989. ; 18 : 664 – 72 . [DOI] [PubMed] [Google Scholar]
- 23. Skiendzielewski JJ, Martyak G . The weak and dizzy patient . Ann Emerg Med 1980. ; 9 : 353 – 6 . [DOI] [PubMed] [Google Scholar]
- 24. NICE . The Guidelines Manual (PMG6 ) . 2012. . http://www.nice.org.uk/article/pmg6/chapter/1%20introduction (accessed on 13 October 2014).
- 25. Price SJ, Shephard EA, Stapley SA, Barraclough K, Hamilton WT . Non-visible versus visible haematuria and bladder cancer risk: a study of electronic records in primary care . BJGP 2014. ; 64 : e584 – 9 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. NICE . The Epilepsies: The Diagnosis and Management of the Epilepsies in Adults and Children in Primary and Secondary Care (CG137 ) . 2012. . http://www.nice.org.uk/guidance/cg137 (accessed on 15 April 2015).
- 27. NICE . Dementia: Supporting People With Dementia and Their Carers in Health and Social Care (CG42 ) . 2006. . http://www.nice.org.uk/guidance/cg42 (accessed on 15 April 2015).
- 28. Hamilton W, Roobottom C . Chronic daily headache . BJGP 2011. ; 61 : 143 – 4 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Elliss-Brookes L, McPhail S, Ives A, et al. . Routes to diagnosis for cancer—determining the patient journey using multiple routine data sets . Br J Cancer 2012. ; 107 : 1220 – 6 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Lyratzopoulos G, Neal RD, Barbiere JM, Rubin GP, Abel GA . 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 : 353 – 65 . [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.