Abstract
Background
Palliative radiotherapy (PRT) comprises half of all radiotherapy use and is an effective and important treatment modality for improving quality of life in incurable cancer patients. We have described the use of PRT in Norway and aimed to identify and quantify the impact of factors associated with PRT utilization.
Material and methods
Population-based data from the Cancer Registry of Norway identified 25,281 patients who died of cancer, 1 July 2009–31 December 2011. Additionally, individual-level data on socioeconomic status and community-level data on travel distance were collected. The proportion of patients who received PRT in the last two years of life (PRT2Y) was calculated, and multivariable logistic regression was used to determine factors that influenced the PRT2Y. Analyses of geographic variation in PRT use were also performed for the time period 2012–2016.
Results
PRT2Y for all cancer sites combined was 29.6% with wide geographic variations (standardized inter-county range; 21.8–36.6%). Female gender, increasing age at death, certain cancer sites, short survival time, and previous receipt of curative radiotherapy were associated with decreased odds of receiving PRT. Patients with low education, those living in certain counties, or with travel distances 100–499 km, were also less likely to receive PRT. Patients with low household income (adjusted odds ratio (OR) = 0.63; 95% confidence interval (CI) = 0.56–0.72) and those diagnosed in hospitals without radiotherapy facility (OR = 0.70; 95% CI = 0.64–0.77) had especially low likelihood of receiving PRT. Significant inter-county variation in use of PRT remained during the time period 2012–2016.
Conclusions
Despite a publicly funded, universal healthcare system with equity as a stated health policy aim, utilization of PRT in Norway is significantly associated with factors such as household income and availability of radiotherapy facility at the diagnosing hospital. Even after adjustments for relevant factors, unexplained geographic variations in PRT utilization exist.
Introduction
Mortality rates in Norway have been quite stable during the last decades, with 10,944 cancer deaths in 2015 [Citation1]. Palliative radiotherapy (PRT) is a time-efficient and cost-effective treatment modality playing an essential role in prevention or management of symptoms in incurable cancer patients. Approximately, half of all radiotherapy courses on a population-basis are prescribed with a palliative intent [Citation2,Citation3] which translates into a minimum need of 1,000,000 PRT courses among Europeans diagnosed with cancer in 2025 (retreatments not included) [Citation4]. Pain, bleeding, obstruction or spinal cord compression-symptoms from primary tumor or metastases, are easily treated with the aim of improving the patients quality of life or survival time [Citation5]. The response rates are usually high with an acceptable tolerance [Citation6–8].
Evidence-based estimates of the population-based demand of PRT in initial cancer treatment exist [Citation9], but similar lifetime estimates are not published. A recently published Canadian study, estimated the optimal PRT rate in a benchmark population with assumed unimpeded radiotherapy access to be 33.9% [Citation10]. Variation in PRT utilization exists in other countries [Citation11–14], and raises concerns whether equal and adequate cancer care is delivered to incurable cancer patients. To date, there are few studies describing PRT utilization in incurable cancer patients for an entire country.
Norway has a scattered population with 5.28 million inhabitants. Publicly financed, universal healthcare is provided, and radiotherapy services are delivered by nine public hospitals with low out-of-pocket expenses for the patient. However, Norwegian radiotherapy rates are documented low compared with evidence-based estimates of optimal use [Citation3]. Under-use of both surgery and radiotherapy was found in lung cancer patients with high age, low education, low income, and those that resided in certain counties [Citation15]. Whether PRT is under-utilized also for the complete cancer population needs further addressing. Better understanding of factors affecting the use of PRT is important to ensure adequate and equal access to this treatment. Few countries are able to use nationwide cancer registry-data with prospectively collected radiotherapy data on each individual cancer case to study these subjects. Several different national registries in Norway provide a unique opportunity to link individual-based data of radiotherapy treatment and demographic data such as socioeconomic factors – enabling accurate estimates in national studies assessing PRT utilization.
The purpose of this nationwide population-based cohort study was to describe the use of PRT in Norway. Further, we aimed to identify and quantify the impact of factors associated with use of PRT and examine whether geographic variation exists.
Material and methods
Data sources and study setting
The Cancer Registry of Norway (CRN) has since 1953 kept a nationwide registry on all malignant neoplasms, with mandatory reporting of all new cancers and all pathology reports of malignancy, ensuring a high degree of data quality and completeness (98.8%) [Citation16]. Since 1997, the CRN has retrieved prospectively registered radiotherapy data on each cancer patient from all Norwegian radiotherapy centers. Linkage to the correct cancer case is made by the unique 11-digit personal identification number and the diagnosis code from the International Classification of Diseases, 10th revision (ICD-10) [Citation3]. The number of radiotherapy centers was stable, while the number of active linear accelerators increased from 39 to 41 during the study-period. The personal identification number also enables data linkage across other nationwide registries. Death certificates are received from the Cause of Death Registry [Citation17], and data on vital status are regularly updated against the National Population Registry. Data on socioeconomic factors for each patient were extracted from the Statistics Norway.
All patients who died of cancer as the underlying cause of death, 1 July 2009–31 December 2011, and who had a diagnosis of cancer registered in the CRN, were eligible for inclusion in this study (n= 26,968). Registered cancer cases based on death certificate or autopsy only (1.6%), and non-melanoma skin cancers (0.4%) were excluded. Recording of radiotherapy data in the CRN started 1 January 1997. Patients diagnosed before this date were therefore excluded – leading to a final study population of 25,287 patients.
Cancer sites were classified based on the ICD-10 codes. For patients with multiple cancer sites (17.5%), the correct cancer case in the CRN to have caused the patient’s death was determined by this prioritized order: (1) complete matching with code of underlying cause of death (ICD-10, 2-digit level); (2) match with underlying cause of death of a cancer within the same organ system [Citation18]; or (3) the latest diagnosed cancer case if no match was found (constituted only 1.4% of the total patient population).
Classification of variables
Radiotherapy variables
Among irradiated patients, 21,285 radiotherapy records were registered. Treatment intent (as prescribed by the oncologist) was registered in 97.2% of these records. For the purpose of this study, radiotherapy records originally categorized as local control (3.6%) were reclassified as palliative. Records missing information on treatment intent (2.8%) were reclassified as curative if the total dose for a radiotherapy course was >39.5 Gy, and palliative if the dose was ≤39.5 Gy. This cutoff value has been considered appropriate for most of the treatment regimes in solid tumors [Citation3,Citation19]. After using this method, the distribution of treatment intention was curative in 28.7%, palliative in 71.2% and unknown in 0.1% of the records.
We used the proportion of patients with at least one record of PRT during the last two years of life (PRT2Y) as an indicator of the lifetime utilization rate of PRT. By following back in time from death patients who had received PRT, we found that 94% had received their last palliative treatment in the last two years of life. The PRT proportion increased very slowly further back in time, and we therefore considered PRT2Y to be a good proxy for lifetime PRT.
Independent variables
Age, municipality/county of residence (at death), gender, and date of diagnosis, radiotherapy treatments, and death, were available in the CRN. ICD-10 diagnoses were combined into 14 diagnostic groups that were used in all analyses. We defined the diagnosing hospital as the hospital where the first malignant biopsy was registered – or, if unknown – the hospital delivering the first clinical report to the CRN. Hospitals were grouped into diagnosing hospital with/without radiotherapy facility. Travel distance was the shortest road distance between the geocoded address of the administrative center in each patients’ municipality and the radiotherapy center where most patients from this municipality were treated (geographic data from The Norwegian Mapping Authority). This distance was calculated using ArcGIS software, performed by the company Geodata AS. The education level achieved at the year of diagnosis was categorized as low (<10 years), middle (10–12 years) and high (≥13 years of education). Data on household income (1 year prior to diagnosis) were available from 2004 and categorized as low (<20th percentile), middle (20th–80th percentile) or high (>80th percentile), within each year.
Geographic variation analyses – time period 2012–2016
By 2007, the planned expansion of new radiotherapy centers in Norway was considered to be complete [Citation3,Citation20] and this together with socioeconomic variables being available up until 31 December 2011, were the reasons for choosing the 2009–2011 study cohort. However, updated Cancer Registry radiotherapy-data on patients who died of cancer during the time period 2012–2016 were available, allowing updated analyses on geographic variation in PRT2Y rates for this time period.
Statistical analyses
We used logistic regression to identify and quantify the impact of factors independently associated with the use of PRT. Variables of potential relevance (based on literature review) were included in two different multivariable models assessing geographic (inter-county) variation of PRT use. Likelihood ratio tests were performed to assess the significance of each individual explanatory variable in the fully adjusted model. Significance levels were set at 5%. Of the variables initially evaluated for inclusion, year of death was the only variable not found to be significant, thus not included in the final models. The base model included county and age at death, gender, cancer site and survival time and was used to study the magnitude of the geographic variation across counties. The fully adjusted model additionally included previous curative radiotherapy, education, household income, radiotherapy facility at diagnosing hospital, and travel distance, and was used to study these factors’ impact on the geographic variation.
Since cause of death information is known to be less accurate for some groups of cancer patients [Citation17], sensitivity analyses were performed where we excluded either patients aged ≥80 years, patients with prostate cancer, or patients with survival time >36 months. The results were similar to the original analyses and made no change in the conclusions. To evaluate the potential bias of missing data, multiple imputation methods [Citation21] were also used to impute missing values in the 15.5% of patients who had one or more missing values for education, household income or diagnosing hospital. Combined results from multivariable regression analyses on 20 imputed datasets were similar to the corresponding analyses on complete case data.
All analyses were performed using Stata, version 14.2 [Citation22]. The Regional Committee for Medical and Health Research Ethics, South Eastern Health Region of Norway, approved this study.
Results
Patient characteristics for the 25,287 patients in the total study cohort are described in . Median age at death was 75 years and median survival time 11.5 months. Median time to death following the last course of PRT was 98 days. Lung cancer patients accounted for 21% of the cohort, colorectal cancer 15%, and other gastrointestinal cancers 15%. More than 50% of the patients lived 50 km or more away from their radiotherapy center.
Table 1. Characteristics of the study population and proportion treated with palliative radiotherapy in the last two years of life (PRT2Y).
Among patients who received PRT, 60% received only one radiotherapy course, 24% received two courses, 9% three courses, 4% four courses and 3% five courses or more.
PRT utilization
The overall proportion of patients who had PRT at least once during the last two years before death, was 29.6% (). Crude PRT2Y rates were highest in patients dying from melanoma, lung, breast, prostate and central nervous system cancer. PRT2Y rates decreased with higher age and was 43% in patients <50 years compared to 17% in those ≥80 years.
PRT2Y varied markedly between the 19 counties. presents the geographic variations in PRT2Y rates standardized to age- and cancer site-distribution in the total study population. The inter-county range in standardized PRT2Y was 21.8% (95% confidence interval (CI): 18.5–25.7%) to 36.6% (95% CI: 32.6–40.8%). Only one of the seven counties with the lowest PRT2Y had a radiotherapy facility within the county, in contrast to six of the seven counties with the highest PRT2Y rates.
Figure 1. Inter-county variation in the proportion of patients who had palliative radiotherapy in the last two years of life (PRT2Y), among patients who died of cancer in Norway, 1 July 2009–31 December 2011. PRT2Y rates are standardized to the age- and cancer site-distribution of the total study population. White dots represent cities with radiotherapy centers.
Factors associated with PRT utilization
lists the results from multivariable logistic regression analysis, which examined the association between independent variables and use of PRT. All factors showed significant association with PRT utilization in both multivariable models. The odds of receiving PRT was negatively associated with female gender and higher age-groups, and was reduced by around 70% in patients aged ≥80 years compared to the youngest patients. Cancer site, survival time and previous treatment with curatively intended radiotherapy, were strongly associated with PRT use. Patients with short education had significantly lower odds of PRT than those with middle, but not with higher education. Patients with low household income had 37% lower odds of PRT than high-income patients. Patients diagnosed at a hospital without radiotherapy facility had 30% lower odds of receiving PRT. Additionally, patients with travel distances between 100 and 500 km were less likely to receive PRT. The effect of travel distance only remained in the group of patients diagnosed at hospitals with no radiotherapy facility, when stratifying on availability of radiotherapy facility at diagnosing hospital (not shown).
Table 2. Odds ratios (OR) of receiving palliative radiotherapy in the last two years of life for patients who died of cancer, 1 July 2009 to 31 December 2011.
Geographic variation in PRT utilization
Compared with the national average PRT2Y, the odds ratios (ORs) for receiving PRT varied considerably across counties in the base model – from 0.63 in the county of Nord-Tr⊘ndelag to 1.48 in the county of Vest-Agder (, base model). After adjustment for patient- and health-system factors not considered to be related to the patients’ needs, significant inter-county variation remained (, fully adjusted model). Supplementary Table 1 and Supplementary Table 2 are full-length versions of and and additionally include county-specific details on PRT2Y rates and ORs, respectively.
Figure 2. Odds ratios for receiving palliative radiotherapy in all counties compared with national average, in patients who died of cancer in Norway, 1 July 2009–31 December 2011. The base model adjusts for gender, age, cancer site and survival time. The fully adjusted model additionally includes previous curative radiotherapy, education, household income, radiotherapy facility at diagnosing hospital, and travel distance.
The national overall PRT2Y rate remained stable (29.8%) for the time period 2012–2016. Significant and wide geographic variation in standardized PRT2Y rates persisted, also after controlling for age, gender, cancer site and survival time differences in logistic regression analyses (Supplementary Table 3).
Discussion
PRT utilization
This population-based cohort study was designed to calculate the proportion of incurable cancer patients treated with PRT, and to identify factors associated with use of PRT. We found that 29.6% of patients who died of cancer July 2009–December 2011, received PRT in the last two years of life. The PRT2Y rates varied widely across the country. The Norwegian PRT2Y rate (all sites combined) is somewhat higher than the PRT2Y of 27.0% for patients who died of cancer in Ontario during 2006–2010, reported by Mackillop and Kong [Citation10]. The same study estimated the optimal lifetime PRT rate and optimal PRT2Y to be 33.9% and 32.4%, respectively (criterion-based benchmarking methodology) – slightly higher than the actual Norwegian PRT2Y. However, these estimates may not be representative for other populations with different age- and cancer-distribution than in Ontario. Based on the results from the multivariable regression analyses, we therefore defined a benchmark population of patients from our study cohort, thought to have optimal access to PRT (patients with high household income and diagnosed in hospitals with radiotherapy facility). The standardized PRT2Y rate for these patients was 37.5% (95% CI: 33.4–41.8%). This rate might serve as a benchmark of optimal PRT rate in Norway – assuming that under-/over-treatment is not a problem in this benchmark population.
Factors associated with PRT utilization
Several factors not assumed related to the patients’ needs, such as gender, age, education, household income, diagnosing hospital with radiotherapy facility, travel distance, and county of residence were independently associated with the use of PRT. Variation in use of PRT related to cancer site is expected, but the significantly higher use in male patients is more difficult to explain. A similar gender-association was documented in some studies [Citation10,Citation12], but was not reconfirmed by others [Citation14,Citation23]. Decreased PRT use among patients with survival time ≤2 months is probably appropriate due to short life expectancy, but significantly decreased PRT2Y even in patients who survived 6–12 months, was more surprising and may reflect underuse of PRT during the last year of life.
Strong associations between older age and low use of PRT are well-documented, also in recent studies [Citation10–15,Citation23]. Our findings may to some extent be confounded by decreased performance status in elderly – a variable not available in our study. However, studies controlling for comorbidity [Citation13,Citation24] and another study assessing the relative decrease in PRT use compared to declining performance status [Citation25], still found significantly decreased PRT rates with higher age. Most of the declining PRT use was explained by decreased referral to cancer centers [Citation25]. A large British survey found older people no more likely to refuse cancer treatment, but actually more likely to trust the recommendations from health professionals than younger patients [Citation26]. This observation challenges the assumption that patient preferences are the cause of declined PRT referral in the elderly, and rather suggests that some patients are not recommended PRT simply due to age alone.
The negative association between previous curatively intended radiotherapy and PRT use is somewhat difficult to interpret. It may reflect lack of referral [Citation12], or maybe refusal, for re-irradiation after recurrence of the primary tumor. Another explanation may be that previous curative radiotherapy has prevented local progression and removed the need of later PRT.
The isolated impact of education level on PRT use has as far as we know, only been described in a Norwegian study on lung cancer [Citation15] with similar results as in our study. The effect of household income was stronger than education in the fully adjusted analysis. The relationship between income and PRT use has been described earlier [Citation11,Citation13,Citation15,Citation23]. However, the strong impact in our study is a worrisome finding in a high-income country with a public healthcare system built on the concept of equity, such as Norway. The underuse of PRT in patients with low socioeconomic status probably reflects that these patients are less likely to seek information or to question their caregivers, if inadequate advices regarding available treatment options are given.
Longer travel distances have been shown to be negatively associated with radiotherapy rates both in palliative [Citation10,Citation12,Citation23] and more general settings [Citation27–30]. Interestingly, although we found a certain association with travel distance, the patients with the longest travel distances tended to have quite high PRT2Y. This raises the question whether use of faster transport alternatives, like aircraft, for long travel distances, may lower the threshold for use of PRT. However, the number of cases was small and differences were non-significant in multivariable analyses. One could argue that travel time may be a better way to evaluate travel distance as a barrier to PRT, but information on the preferred type of transportation from different regions in the country was not easily accessible.
The strong relation between availability of radiotherapy facility in diagnosing hospitals and high PRT use was also found in Canadian studies [Citation10,Citation11,Citation23] and showed higher impact than travel distance on PRT use in our study. Lack of knowledge of common indications or benefits of PRT, and uncertainty about the referral process among general practitioners, is documented to influence the referral of patients for PRT [Citation31,Citation32]. Our finding suggests that this probably also applies to specialists in hospitals without radiotherapy facilities where such information may be less available.
Geographic variation in PRT utilization
Even after controlling for all the factors noted above, relatively large geographic differences in utilization of PRT remain unexplained. Significant and wide geographic variation in standardized PRT2Y rates persisted during 2012–2016.
A limitation of our study is the lack of data on factors that could provide a more complete picture of radiotherapy practice patterns across Norwegian counties. Long waiting times are reported to be a perceived potential barrier to radiotherapy referral by health professionals [Citation32,Citation33], but data on this factor were not available. Treatment center differences in use of single fraction regimens for bone metastases were found in one Norwegian study [Citation34] and the relation to total use of PRT is of interest as treatment schedule differences are shown to influence the caregivers decisions to refer patients [Citation35]. Unfortunately, information regarding irradiated anatomical region was only easily available from some hospitals for the study period 2007–2011, thus this question could not be examined properly. Our findings may also reflect possible geographical differences in oncologists’ management preferences regarding chemotherapy/other tumor-targeted treatment alternatives, opioid-prescription [Citation36], or simply just different thresholds for accepting PRT referrals. For instance, substantial variation in the management of stage IV lung cancers was found among radiation oncologists in an American study [Citation37]. Geographical differences in density of oncologists in Norway exist, but were not explored in our study.
The major strength of our study is the use of the national high-quality registries for all variables included in this population-based study of patients who died of cancer in Norway. Radiotherapy data and most other variables were individual-based – ensuring high accuracy in our estimates. Despite limited degree of missing data, multiple imputation analyses were performed, but changed no results or conclusions.
Conclusions
The publicly funded health care system in Norway is built on the concept of equity in access to necessary health services for patients with equal medical needs. Still, this population-based study found substantial variations in the use of PRT across Norway. Factors unrelated to patients’ needs such as, advanced age, low household income, county of residence, diagnosing hospitals without radiotherapy facility, and longer travel distances, were all strongly associated with decreased likelihood of receiving PRT.
Possible geographical differences among different radiotherapy centers concerning treatment capacity/oncologist staffing, treatment practice, and use of ambulatory oncology care, should be further explored as potential barriers to PRT. There is a need to improve awareness and knowledge of indications and benefits of PRT on the patient and referrer level to ensure equal and adequate palliative cancer care in Norway.
Linn_M.__sli_et_al._Tables_Supplementary_Tables_1-3.docx
Download MS Word (42.7 KB)Disclosure statement
No potential conflict of interest was reported by the authors.
Additional information
Funding
References
- Cancer Registry of Norway. Cancer in Norway 2016 – cancer incidence, mortality, survival and prevalence in Norway. Oslo: Cancer Registry of Norway; 2017.
- Moller TR, Brorsson B, Ceberg J, et al. A prospective survey of radiotherapy practice 2001 in Sweden. Acta Oncol. 2003;42:387–410.
- Åsli LM, Kvaløy SO, Jetne V, et al. Utilization of radiation therapy in Norway after the implementation of the national cancer plan: a national, population-based study. Int J Radiat Oncol Biol Phys. 2014;90:707.
- Borras JM, Lievens Y, Barton M, et al. How many new cancer patients in Europe will require radiotherapy by 2025? An ESTRO-HERO analysis. Radiother Oncol. 2016;119:5–11.
- Lutz ST, Jones J, Chow E. Role of radiation therapy in palliative care of the patient with cancer. J Clin Oncol. 2014;32:2913–2919.
- Cameron MG, Kersten C, Vistad I, et al. Palliative pelvic radiotherapy for symptomatic rectal cancer – a prospective multicenter study. Acta Oncol. 2016;55:1400–1407.
- Cameron MG, Kersten C, Vistad I, et al. Palliative pelvic radiotherapy for symptomatic incurable prostate cancer – a prospective multicenter study. Radiother Oncol. 2015;115:314–320.
- Lutz S, Balboni T, Jones J, et al. Palliative radiation therapy for bone metastases: update of an ASTRO Evidence-Based Guideline. Pract Radiat Oncol. 2017;7:4–12.
- Jacob S, Wong K, Delaney GP, et al. Estimation of an optimal utilisation rate for palliative radiotherapy in newly diagnosed cancer patients. Clin Oncol (R Coll Radiol). 2010;22:56–64.
- Mackillop WJ, Kong W. Estimating the need for palliative radiation therapy: a benchmarking approach. Int J Radiat Oncol Biol Phys. 2016;94:51–59.
- Huang J, Zhou S, Groome P, et al. Factors affecting the use of palliative radiotherapy in Ontario. Clin Oncol. 2001;19:137.
- Lavergne MR, Johnston GM, Gao J, et al. Variation in the use of palliative radiotherapy at end of life: examining demographic, clinical, health service, and geographic factors in a population-based study. Palliat Med. 2011;25:101–110.
- Murphy JD, Nelson LM, Chang DT, et al. Patterns of care in palliative radiotherapy: a population-based study. J Oncol Pract. 2013;9:e220–e227.
- Huang J, Wai ES, Lau F, et al. Palliative radiotherapy utilization for cancer patients at end of life in British Columbia: retrospective cohort study. BMC Palliat Care. 2014;13:49.
- Nilssen Y, Strand TE, Fjellbirkeland L, et al. Lung cancer treatment is influenced by income, education, age and place of residence in a country with universal health coverage. Int J Cancer. 2016;138:1350–1360.
- Larsen IK, Smastuen M, Johannesen TB, et al. Data quality at the Cancer Registry of Norway: an overview of comparability, completeness, validity and timeliness. Eur J Cancer. 2009;45:1218–1231.
- Skyrud KD, Bray F, Moller B. A comparison of relative and cause-specific survival by cancer site, age and time since diagnosis. Int J Cancer. 2014;135:196–203.
- Howlader N, Ries LA, Mariotto AB, et al. Improved estimates of cancer-specific survival rates from population-based data. J Natl Cancer Inst. 2010;102:1584–1598.
- Mackillop WJ, Zhou S, Groome P, et al. Changes in the use of radiotherapy in Ontario 1984–1995. Int J Radiat Oncol Biol Phys. 1999;44:355–362.
- Ministry of Health and Care Services. Omsorg og kunnskap! Norsk kreftplan. NOU 1997: 20 [Care and knowledge! The Norwegian Cancer Plan. NOU 1997: 20]. Oslo: Ministry of Health and Care Services; 1997.
- White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues and guidance for practice. Stat Med. 2011;30:377–399.
- StataCorp. Stata Statistical Software: Release 14. College Station (TX): StataCorp LP; 2015.
- Sutton DS, Kong W, Ding K, et al. The use of palliative radiotherapy for bone metastasis. Radiother Oncol. 2010;97:548–553.
- Wong J, Xu B, Yeung HN, et al. Age disparity in palliative radiation therapy among patients with advanced cancer. Int J Radiat Oncol Biol Phys. 2014;90:224–230.
- Tyldesley S, Zhang-Salomons J, Groome PA, et al. Association between age and the utilization of radiotherapy in Ontario. Int J Radiat Oncol Biol Phys. 2000;47:469–480.
- IPSOS MORI. Older people living with cancer. Exploring the attitudes and behaviours of older people living with cancer. London (UK): Macmillan Cancer Support; 2015. [Internet] Available from: http://www.macmillan.org.uk/Documents/Campaigns/AttitudesofolderpeoplelivingwithcancerAugust15.pdf
- Hanna TP, Richardson H, Peng Y, et al. A population-based study of factors affecting the use of radiotherapy for endometrial cancer. Clin Oncol (R Coll Radiol). 2012;24:e113–e124.
- Tyldesley S, McGahan C. Utilisation of radiotherapy in rural and urban areas in British Columbia compared with evidence-based estimates of radiotherapy needs for patients with breast, prostate and lung cancer. Clin Oncol (R Coll Radiol). 2010;22:526–532.
- Gabriel G, Barton M, Delaney GP. The effect of travel distance on radiotherapy utilization in NSW and ACT. Radiother Oncol. 2015;117:386–389.
- Liu E, Santibanez P, Puterman ML, et al. A quantitative analysis of the relationship between radiation therapy use and travel time. Int J Radiat Oncol Biol Phys. 2015;93:710–718.
- Vulto A, van Bommel M, Poortmans P, et al. General practitioners and referral for palliative radiotherapy: a population-based survey. Radiother Oncol. 2009;91:267–270.
- Samant RS, Fitzgibbon E, Meng J, et al. Barriers to palliative radiotherapy referral: a Canadian perspective. Acta Oncol. 2007;46:659–663.
- Sundaresan P, Milross CG, Smith A, et al. Factors influencing the use of RT in NSW: a qualitative study exploring consumer and health professional perspectives. J Med Imaging Radiat Oncol. 2014;58:625–632.
- Laugsand TS, Kaasa S, Romundstad P, et al. Radiotherapy for bone metastases: practice in Norway 1997–2007. A national registry-based study. Acta Oncol. 2013;52:1129–1136.
- Saito T, Toya R, Semba A, et al. Influence of the treatment schedule on the physicians' decisions to refer bone metastases patients for palliative radiotherapy: a questionnaire survey of physicians in various specialties. Nagoya J Med Sci. 2016;78:275–284.
- Olsen AS, Ottesen S. Variable prescription of opioids to cancer patients in Norway. Tidsskr nor Laegeforen. 2008;128:1271–1274.
- Kong FM, Cuneo KC, Wang L, et al. Patterns of practice in radiation therapy for non-small cell lung cancer among members of the American Society for Radiation Oncology. Pract Radiat Oncol. 2014;4:e133–e141.