Radiotherapy and Oncology 151 (2020) 200–205 Contents lists available at ScienceDirect Radiotherapy and Oncology journal homepage: www.thegreenjournal.com Original Article Impact of non-adherence to radiotherapy on 1-year survival in cancer patients in Catalonia, Spain Josep M. Borras a,⇑, Rebeca Font b, Judit Solà b, Miquel Macia c, Victòria Tuset d, Meritxell Arenas e,f, Arantxa Eraso g, Ramona Verges h, Nuria Farré i, Agustin Pedro i, Meritxell Mollà j, Manel Algara k, Josep M. Solé l, Moises Mira m, Josep A. Espinàs b a Department of Clinical Sciences, University of Barcelona and Bellvitge Biomedical Research Institute (IDIBELL); b Cancer Strategy, Department of Health, Barcelona; c Radiation Oncology Department, Institut Català d’Oncologia, Hospital Duran i Reynals, L’Hospitalet de Llobregat; d Radiation Oncology Department, Institut Català d’Oncologia, Hospital Germans Trias i Pujol, Badalona; e Radiation Oncology Department, Hospital Universitari Sant Joan de Reus; f Faculty of Medicine, University of Rovira i Virgili; g Radiation Oncology Department, Institut Català d’Oncologia, Hospital Trueta, Girona; h Radiation Oncology Department, Hospital Universitari de la Vall d’Hebron; i Radiation Oncology Department, Hospital de la Santa Creu i Sant Pau, Barcelona; j Department of Radiation Oncology, Hospital Clinic de Barcelona; k Radiation Oncology Department, Hospital del Mar, Radiation Oncology Research Group, IMIM and Universitat Autònoma de Barcelona; l Radiation Oncology Department, Consorci Sanitari Terrassa- Hospital General Catalunya- Hospital Manresa, Barcelona; and m Radiation Oncology Department, H.U. Arnau de Vilanova, Lleida, Spain a r t i c l e i n f o a b s t r a c t Article history: Background: This study aims to assess the effects of non-adherence to external beam radiation therapy in Received 27 February 2020 cancer patients receiving treatment with a curative. Received in revised form 23 July 2020 Methods: This retrospective cohort study collected health records data for all cancer patients treated with Accepted 1 August 2020 external beam radiotherapy with curative intent in 2016 in Catalonia, Spain. Adherence was defined as Available online 6 August 2020 having received at least 90% of the total dose prescribed. A logistic regression model was used to assess factors related to non-adherence, and its association with one-year survival was evaluated using Cox Keywords: regression. Adherence External radiotherapy Results: The final sample included 8721 patients (mean age 63.6 years): breast cancer was the most com- Dose mon tumour site (38.1%), followed by prostate and colon/rectum. Treatment interruptions prolonged the Survival total duration of therapy in 70.7% of the patients, and 1.0% were non-adherent. Non-adherence was asso- ciated with advanced age, female gender, and some localization of primary tumour (head and neck, uri- nary bladder, and haematological cancers). The risk of death in non-adherent patients was higher than in adherent patients (hazard ratio [HR] 1.63, 95% confidence interval 0.97–2.74), after adjusting for the potential confounding effect of age, gender, tumour site and comorbidity. Conclusion: Non-adherence to radiotherapy, as measured by the received dose, is very low in our setting, and it may have an impact on one-year survival. Ó 2020 The Authors. Published by Elsevier B.V. Radiotherapy and Oncology 151 (2020) 200–205 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Adherence to chemotherapy and hormone therapy has proven mone therapy during their first year of treatment at 79%, dropping to be a relevant problem in cancer patients undergoing different to 56% in the fifth year [7]. treatment regimens [1–3]. Adherence is defined as the degree to Radiotherapy differs from oral endocrine therapy and which the patient’s behaviour is consistent with the therapy pre- chemotherapy in that it generally has a short duration (maximum scribed by their doctor, and normally it is measured as the percent- 8 weeks) and entails regular medical and technical supervision age of doses received relative to those indicated. The most typical over the course of treatment. This context is favourable to high cutoff to define adherence in these terms is 80% [4]. In patients adherence rates once the patient has agreed to undergo treatment. with breast cancer, non-adherence is associated with a higher risk As a result, assessing adherence has not been considered very rel- of recurrence and death [5,6]. One literature review estimated the evant in radiation oncology, and what research has been published proportion of patients with breast cancer who are adherent to hor- is oriented toward assessing interruptions (missed appointments during the course of treatment) rather than evaluating adherence in terms of receiving the prescribed doses [8,9]. Since the 1990s, the public healthcare system in Catalonia ⇑ Corresponding author at: Department of Clinical Sciences, University of (Spain) has maintained a registry of all patients receiving radio- Barcelona, Gran via 199, 089908-Hospitalet, Barcelona, Spain. therapy with public financing. This study makes use of these data, E-mail address: jmborras@ub.edu (J.M. Borras). https://doi.org/10.1016/j.radonc.2020.08.002 0167-8140/Ó 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). J.M. Borras et al. / Radiotherapy and Oncology 151 (2020) 200–205 201 aiming to assess adherence to external radiotherapy measured by patients were excluded from the analysis. We also excluded from the percentage of the total dose received in all cancer patients trea- this analysis patients under the age of 18, those receiving only ted with radiotherapy and the impact on one-year survival. brachytherapy, patients with treatment interruptions for medical reasons (poor general condition due to treatment, disease progres- Material and methods sion, intercurrent illness, and/or treatment change), and those who died within the first month of finishing treatment. We considered The initial cohort included all cancer patients indicated for these to be cases with an inappropriate indication for radiotherapy with a curative intent. treatment with radiotherapy in 2016, provided they had received at least one fraction of the prescribed treatment. We collected ret- Adherence was calculated as the percentage of doses received relative to those prescribed. The cut-off was set at 90%, in agree- rospective data from the register of patients treated with external ment with clinicians and taking into account the type of treatment beam radiotherapy in the 11 publicly funded radiation oncology under study (short duration) and its therapeutic importance. In our services in Catalonia. This represents about 90% of all patients care context, the radiation oncology services have protocols in receiving radiation therapy in the region. place to calculate the necessary doses according to the service- The variables included were: hospital, age, gender, tumour site related, logistical, or schedule-related interruptions experienced, targeted, treatment indication (curative versus palliative), total doses and sessions prescribed, final doses and sessions received, and to extend the length of treatment in order to ensure that patients receive at least the total doses prescribed. Patients who date of treatment initiation and finalization, and treatment inter- ruptions and their causes. The causes of interruption considered did not receive 90% of the total dose due to non-medical reasons, including patient-related ones, were considered non-adherent. included those related to the equipment (both unexpected break- downs and planned maintenance); logistical challenges affecting We performed a descriptive statistical analysis of the variables and a multivariable logistic regression, using adherence as the out- patients (usually transport); public holidays; medical problems related to the patient (disease progression, intercurrent illness, come variable and adjusting for hospital, age, gender, comorbidity and tumour site. Results are expressed as odds ratios (OR) with and/or treatment change); and personal reasons. It was possible to consider more than one cause of treatment interruption in the their 95% confidence intervals (CI). For the survival analysis, we constructed a Kaplan-Meier curve and fit a Cox proportional haz- analysis. The minimum basic discharge data set was the source of data to identify the comorbidity (Charlson Comorbidity Index) ards model, adjusting for age, sex, hospital, comorbidity and the diagnosis motivating the radiotherapy indication. All analyses of patients before starting radiotherapy treatment. Vital status was ascertained from the Catalan Health Service database of were undertaken using SPSS software (version 21). insured persons, which is updated monthly with the official mor- tality statistics. For reasons of confidentiality, it was not possible Results to obtain data on the specific cause of death. The analysis was restricted to the patients receiving treatment Fig. 1 presents the patient selection flowchart. Of the 15,501 with a curative intent, defined from the standard prescribed ses- patients treated, 371 were excluded because they did not meet sions according to tumour site. This number of sessions was estab- the inclusion criteria, 5133 because they received radiotherapy lished for each tumour site; this definition allowed for some with a palliative intent, 728 because the patients stopped treat- patients with metastasis to be included in the cohort, but these ment for medical reasons, 446 because they had metastasis and Fig. 1. Flow chart of patients. 202 Impact of non-adherence to radiotherapy on survival 102 who died within a month of finalizing treatment. The final Adherence is probably of special importance for tumour sites cohort was 8721 patients: 38.1% had breast cancer; 16.5%, prostate such as the head and neck, bladder and cervix, as well as haemato- cancer; and 9.2%, colorectal cancer. Smaller proportions had can- logical cancers, which in our setting were all associated with a cers of the head and neck, the lung, and other types. higher risk of non-adherence. It is possible that these patients Table 1 shows the patient characteristics. Over half were dropped out of treatment for reasons that are not contemplated women, and their mean age was 63.6 years. One per cent were in our study, for example socioeconomic determinants [11]. In non-adherent. Treatment interruptions for any reason affected studies performed in populations with low socioeconomic status, 70.7% of the patients with at least one missed appointment. Most tumours of the cervix and the head and neck were also associated of the interruptions were related to equipment failures/mainte- with a greater risk of missed appointments in multivariable analy- nance and public holidays. Just 4.8% of the treatments were inter- sis [8]. Elsewhere, too, head and neck cancers were associated with rupted for personal reasons, while 13.0% of the interruptions were a greater risk of missed appointments, as reported by Rangarajan in for other reasons. Interruptions extended treatment duration by India [12]. more than two days in 69.1% of the cases (6020/8712). Also, In radiation oncology, a large and consistent body of evidence 45.6% of de patients had no comorbidities. shows that treatment interruptions for any reason worsen the Table 2 presents the factors related to non-adherence (less than patient’s prognosis [11,13–17]. This knowledge underpins the 90% of the prescribed doses received). Women showed similar notable efforts made to compensate interruptions with additional adherence than men (OR 1.19 95% CI 0.63–2.24). Advanced age fractions in order to administer at least the prescribed doses by appeared to confer a higher risk of non-adherence; however, nei- treatment end. The number of interruptions due to weekday public ther the results for gender nor age were statistically significant. holidays is very high in our country. This factor and others mean Some diagnoses did show a significant association with non- adherence, including tumours of the head and neck (OR 2.64 95% CI 1.01–6.91). Patients with haematological cancers (OR 2.80 95% Table 1 Description of patient characteristics (N = 8721*). CI 0.290–8.76) and cervical cancer (OR 2.08 95% CI 0.47–9.09) also tended to be less adherent. On the other hand, patients with breast n % and prostate cancers showed high levels of adherence. Non adher- Gender ence increased with number of comorbid pathologies, although Men 3955 45.4 this association was not statistically significant. Women 4766 54.6 Table 3 presents the results of the Cox model. After adjusting for Age (mean 63.6 ± 13.5 years) age, gender, hospital, comorbidity and diagnosis, there was a not 49 years 1298 14.9 50–59 years 1843 21.1 statistically significant difference in survival between patients 60–69 years 2368 27.2 who did and did not receive at least 90% of the planned treatments. 70–79 years 2250 25.8 Non-adherent patients had a higher risk of death at one year com- 80 years 962 11.0 pared to adherent ones (OR 1.63 95% CI 0.97–2.74). Fig. 2 shows Tumour site the survival curve according to adherence, as defined by percent- Head and neck 714 8.2 age of the dose received (excluding patients who stopped treat- Colorectal 798 9.2 Trachea, bronchus, lung 581 6.7 ment for medical reasons). When the analysis focused on Skin 132 1.5 patients who dropped out for exclusively personal reasons, the risk Breast 3322 38.1 of mortality was also double, although given the low volume of Bones and connective tissues 129 1.5 cases (only 291 patient dropped out exclusively for patient- Uterus 178 2.0 related reasons) the results did not reach statistical significance Prostate 1436 16.5 Urinary bladder 80 0.9 after adjusting for age, gender, tumour site and comorbidity (OR Central nervous system 303 3.5 1.88 95% CI 0.41–8.61). Haematological 305 3.5 Cervix 132 1.5 Discussion Other digestive 326 3.7 Other 385 3.3 The proportion of patients who did not adhere to radiotherapy Concomitant chemotherapy Yes 2202 25.2 (as measured by those not receiving a sufficient percentage of their No 6218 71.3 prescribed doses) was very low compared to that observed in Unknown 301 3.5 patients receiving other oncological treatments, like oral Comorbidities chemotherapy or hormone therapy. Just 1.0% of all 8721 included Non 3978 45.6 patients receiving radiotherapy with curative intent in 2016 in Cat- 1–2 1771 20.3 alonia suspended their course of treatment before reaching 90% of 3+ 432 5.0 the total doses prescribed. When non-adherence was attributable Unknown 2540 29.1 Death at one year 824 9.8 to non-medical reasons (progression of the disease, intercurrent Non-adherent 87 1.0 disease and/or change of therapy), it was associated with lower Total treatment interruptions due to: 6170 70.7 one-year survival. This result clearly indicates that radiotherapy Equipment 3884 62.9 is associated with high rates of adherence in our setting, thanks Personal reasons 297 4.8 Logistical reasons 805 13.0 in part to its short duration and the strict medical supervision Public holidays 5191 84.1 involved. Classically, shorter treatments are associated with a Days of treatment prolongation lower risk of non-adherence [10]. 2 days 2692 30.9 This result implies two relevant corollaries: first, non- 3–4 days 2509 28.8 adherence to radiotherapy is associated with an increased risk of 5–7 days 2051 23.5 death in cancer patients receiving this therapy, and second, there 8–9 days 613 7.0 are some factors that can slightly increase the risk of non- 10 days 847 9.7 adherence, including advanced age and diagnosis of some tumour * Total number of cases N = 8721. The difference relative to the total N corre- types. sponds to missing values. J.M. Borras et al. / Radiotherapy and Oncology 151 (2020) 200–205 203 Table 2 Prognostic and therapeutics factors associated with non-adherence. Non-adherence N (% Non-adherent) ORa (95% CI) p Adherent 8595 1.0 Gender Men 3890 1.1 1 Women 4705 0.9 1.19 (0.63–2.24) 0.589 Age group 49 years 1285 1.1 1 50–59 years 1820 1.0 1.05 (0.51–2.16) 0.898 60–69 years 2327 0.6 0.61 (0.28–1.33) 0.212 70–79 years 2218 1.1 1.12 (0.54–2.36) 0.759 80 years 945 1.6 1.37 (0.61–3.07) 0.448 Tumour site Colorectal 784 0.8 1 Head and neck 695 2.3 2.64 (1.01–6.91) 0.048* Trachea, bronchus, lung 570 1.1 1.35 (0.43–4.29) 0.609 Skin 130 2.3 2.42 (0.58–10.12) 0.225 Breast 3283 0.7 0.77 (0.28–2.09) 0.608 Bones & connective tissues 129 2.3 2.56 (0.60–10.86) 0.203 Uterus 172 1.2 1.13 (0.21–5.99) 0.888 Prostate 1418 0.8 1.10 (0.39–3.07) 0.863 Urinary bladder 79 3.8 4.69 (1.08–20.24) 0.039* Central nervous system 293 0.7 0.70 (0.14–3.60) 0.671 Haematological 305 2.3 2.80 (0.90–8.76) 0.077 Cervix 132 2.3 2.08 (0.47–9.09) 0.333 Other digestive 324 0.0 - 0.994 Other 281 0.7 0.83 (0.16–4.22) 0.818 Comorbidities 0 3913 0.8 1 1–2 1733 1.2 1.30 (0.72–2.36) 0.380 3+ 424 1.7 1.71 (0.72–4.07) 0.224 Missing 2525 1.1 1.34 (0.80–2.25) 0.268 Interruptions No 2542 1.3 1 Yes 6053 0.9 1.32 (0.76–2.29) 0.323 Causes of interruption Equipment No 2278 1.4 1 Yes 3775 0.6 0.58 (0.32–1.07) 0.081 Personal reasons No 5756 0.5 1 Yes 297 9.4 19.65 (10.19–37.87) <0.001* Logistical reasons No 5248 0.9 1 Yes 805 1.1 1.79 (0.71–4.55) 0.219 Public holidays No 950 2.9 1 Yes 5103 0.5 0.19 (0.11–0.35) <0.001* Death at one year No 7493 0.9 1 Yes 806 1.9 1.82 (0.96–3.44) 0.066 n (%): number of cases (% non-adherence); The difference relative to the total N corresponds to missing values. ORa: odds ratio adjusted for centre, age group, gender, diagnosis, Comorbidities; CI: confidence interval. * Statistical significance, P < 0.05. that only about a third of the treatments in our setting finish by the 12.7% of the cases, this kept patients from receiving their full dose, expected date, necessitating a substantial effort in clinical manage- and a quarter saw their treatment duration extended by more than ment to complete the prescribed doses on time. The success of five days [18]. Thus, adherence can be conceived in two ways: as these efforts is evidenced by the low number of patients who do missed appointments and as the proportion of doses received rel- not receive their full dose once treatment has started. Thus, even ative to those prescribed. With some exceptions [18], researchers the patients whose treatment duration is extended by more than have focused on the former, analysing the dose received only in a two days have a high probability of finally receiving the totality complementary way. In contrast, our study sheds light on the of the prescribed dose, with the subsequent benefits on their prog- impact of doses on the patient’s prognosis, which should be anal- nosis. There is another option to cope with this problem, namely, ysed independently from the prolongation of the treatment dura- adding a second fraction on some treatment days, which was the tion, even though its quantitative impact is limited. In that sense, option chosen in 12.1% of the cases in our cohort. Another option our proposal builds on Khalil’s work studying the impact of the would be to work on Saturdays, as in other countries, which would total dose lost as a necessary measure to complement the data solve both problems (doses and overall treatment time), but this on missed appointments. Indeed, the effectiveness of the treatment option is not easily applicable in our health system due to manage- is a function of both the total doses administered and the treat- ment criteria for organizing of delivery of radiation oncology. ment time and fractioning. These results also pose a problem related to measuring adher- Another aspect to highlight is that in this project, we made a ence in radiation oncology, as practically all the published litera- considerable effort to specify the causes of the treatment interrup- ture uses missed appointments, classified with criteria similar to tion. In some cases, treatments were interrupted for more than one ours, to study this endpoint. In fact, even in a clinical trial per- reason, and in our analysis, we excluded the patients whose treat- formed in patients with high-risk tumours, like those of the head ment was interrupted for medical reasons upon careful review by and neck, the probability of missed appointments was very high, the attending clinicians. We considered only patients who were affecting up to 70% of the included patients. In an estimated non-adherent for other reasons, and this decision could explain 204 Impact of non-adherence to radiotherapy on survival Table 3 Cox regression for one-year survival. Survival at one year of treatment end n (% died) HRa (95% CI) p Death at one year of treatment end 8299 9.7 Adherence Adherent 8216 9.6 1 Non-adherent 83 18.1 1.63 (0.97–2.74) 0.066 Gender Male 3808 14.2 1 Female 4608 6.2 0.93 (0.78–1.11) 0.416 Age group 49 years 1243 6.3 1 50–59 years 1775 7.0 1.06 (0.79–1.43) 0.690 60–69 years 2286 9.5 1.40 (1.06–1.85) 0.017* 70–79 years 2185 10.7 1.80 (1.36–2.38) <0.001* 80 years 927 18.6 2.64 (1.96–3.56) <0.001* Tumour site Colorectal 765 6.5 1 Head and neck 681 18.9 3.40 (2.43–4.77) <0.001* Trachea, bronchus, lung 561 32.3 5.44 (3.94–7.52) <0.001* Skin 129 24.0 2.69 (1.69–4.28) <0.001* Breast 3223 1.7 0.31 (0.20–0.47) <0.001* Bones and connective tissues 124 14.5 2.59 (1.47–4.56) <0.001* Uterus 170 5.9 0.79 (0.37–1.70) 0.551 Prostate 1397 1.3 0.17 (0.09–0.29) <0.001* Urinary bladder 79 44.3 5.63 (3.59–8.82) <0.001* Central nervous system 291 35.1 7.98(5.56–11.43) <0.001* Haematological 286 8.7 1.50 (0.92–2.45) 0.104 Cervix 127 15.7 3.07 (1.79–5.27) <0.001* Other digestive 311 33.4 6.25 (4.42–8.86) <0.001* Other 272 16.5 2.70 (1.77–4.10) <0.001* Comorbidities 0 3953 7.8 1 1–2 1763 14.7 1.15 (0.96–1.36) 0.128 3+ 431 21.3 1.35 (1.05–1.72) 0.017* Missing 2269 7.3 0.90 (0.74–1.09) 0.273 n (%): number of cases (% of death patients at one year of treatment end). The difference relative to the total N corresponds missing values. HRa: hazard ratio adjusted for centre, age group, gender, diagnosis, comorbidities; CI: confidence interval. *Statistical significance, P < 0.05. Fig. 2. Overall survival by adherence. J.M. Borras et al. / Radiotherapy and Oncology 151 (2020) 200–205 205 why our results showed a much lower proportion of non-adherents Funding (1.0%) than other studies, for example, Khalil’s [18], who reported that 12.7% of patients with head and neck cancer were non- We thank CERCA Programme/Generalitat de Catalunya for adherent. institutional support (AGAUR 2017SGR735). This study has been Some limitations of this study should be mentioned. First, due funded by the project PI15/00945, (Co-funded by European to reasons of confidentiality, it was not possible to ascertain the Regional Development Fund. ERDF, a way to build Europe). cause of death, precluding an analysis of cause-specific survival, which would have been more appropriate. Second, in 3.5% of the References cases it was not possible to ascertain the vital status at one year after treatment. Third, although we adjusted for comorbidity, in [1] Darkow T, Henk HJ, Thomas SK, Feng W, Baladi JF, Goldberg GA, et al. Treatment interruptions and non-adherence with imatinib and associated 29.1% of the cases it was not possible to determine this information healthcare costs: a retrospective analysis among managed care patients with prior to the radiation oncology treatment, so we opted to include chronic myelogenous leukaemia. Pharmacoeconomics 2007;25:481–96. the patients with missing information for this variable as a cate- [2] Mathes T, Pieper D, Antoine S-L, Eikermann M. Adherence influencing factors in patients taking oral anticancer agents: a systematic review. Cancer gory in the multivariate analysis. Fourth, no data were available Epidemiol 2014;38:214–26. on stage, either at diagnosis or before treatment initiation. Lastly, [3] Font R, Espinas JA, Layos L, Martinez Villacampa M, Capdevila J, Tobeña M, et al. we were not able to analyse some other variables of interest, such Adherence to capecitabine in preoperative treatment of stage II and III rectal cancer: do we need to worry?. Ann Oncol 2017;28:831–5. as socioeconomic status or place of residence, which may be [4] Cramer JA, Roy A, Burrell A, Fairchild CJ, Fuldeore MJ, Ollendorf DA, et al. important explanations for non-adherence. Medication compliance and persistence: terminology and definitions. Value One strength of this study is the fact that it is population-based Health 2008;11:44–7. rather than limited to patients involved in a clinical trial for speci- [5] Hershman DL, Shao T, Kushi LH, Buono D, Tsai WY, Fehrenbacher L, et al. Early discontinuation and non-adherence to adjuvant hormonal therapy are fic diagnoses. This approach minimized the selection bias, although associated with increased mortality in women with breast cancer. Breast we did not included the estimated 10% of patients receiving treat- Cancer Res Treat 2011;126:529–37. ment in private facilities. The definition of the study population as [6] Font R, Espinas JA, Barnadas A, Izquierdo A, Galceran J, Saladie F, et al. Influence of adherence to adjuvant endocrine therapy on disease-free and overall cases treated with a curative intent, and the detailed definition of survival: a population-based study in Catalonia, Spain. Breast Cancer Res Treat the causes of treatment interruption, are other relevant aspects, 2019;175:733–40. as is the review of dubious cases by clinicians. Also, the small per- [7] Moon Z, Moss-Morris R, Hunter MS, Carlisle S, Hughes LD. Barriers and facilitators of adjuvant hormone therapy adherence and persistence in women centage of cases with metastasis at the start of treatment were with breast cancer: a systematic review. Patient Prefer Adher 2017;11:305–22. excluded, even though treatment intent was defined as curative, [8] Ohri N, Rapkin BD, Guha D, Haynes-Lewis H, Guha C, Kalnicki S, et al. Predictors to minimize bias. Palliative indications could likewise have influ- of radiation therapy noncompliance in an urban academic cancer center. Int J Radiat Oncol Biol Phys 2015;91:232–8. enced non-adherence, possibly reducing it further due to shorter [9] Shi DD, Connolly RJ, Shih HA. Understanding causes for missing radiation treatments [19], although these patients have limited survival treatments as a step towards improving adherence and quality of patient care. time, which would have made it very difficult to pool these Int J Radiation Oncol Biol Phys 2017;99:E415. [10] DiMatteo R. Variations in patient adherence to medical recommendations: a patients in an analysis of those treated with curative intent due quantitative review of 50 years of research. Med Care 2004;42:200–9. to the different aims of the therapy. [11] Arenas M, Sabater S, Gascón M, Henríquez I, Bueno MJ, Rius À, et al. Quality Radiation oncology is a basic pillar of multidisciplinary cancer assurance in radiotherapy: analysis of the causes of not starting or early treatment, and it contributes independently to local disease con- radiotherapy withdrawal. Radiat Oncol 2014;9. [12] Rangarajan R, Jayaraman K. Barriers affecting adherence to radiation treatment trol and overall survival [20]. It is necessary to measure non- and strategies to overcome those barriers. Indian J Cancer 2017;54:458. adherence based on total doses received relative to prescribed [13] Bese NS, Hendry J, Jeremic B. Effects of prolongation of overall treatment time doses, not just based on missed appointments. The percentage of due to unplanned interruptions during radiotherapy of different tumor sites and practical methods for compensation. Int J Radiat Oncol Biol Phys non-adherent patients is very low (1.0%) according to the dose 2007;68:654–61. received, whereas 70.7% of the treatments had interruptions due [14] Fesinmeyer MD, Mehta V, Blough D, Tock L, Ramsey SD. Effect of radiotherapy to weekday public holidays, machine malfunctions, equipment interruptions on survival in medicare enrollees with local and regional head- and-neck cancer. Int J Radiat Oncol Biol Phys 2010;78:675–81. maintenance, and personal reasons. Considering the quantity of [15] Thames HD, Kuban D, Levy LB, Horwitz EM, Kupelian P, Martinez A, et al. The interruptions due to public holidays, expanding services to Satur- role of overall treatment time in the outcome of radiotherapy of prostate days should be considered, as this could minimize the impact on cancer: An analysis of biochemical failure in 4839 men treated between 1987 and 1995. Radiother Oncol 2010;96:6–12. the patient and allow compensation for total doses and treatment [16] Sanpaolo P, Barbieri V, Genovesi D. Biologically effective dose and definitive time. Overall, non-adherence may have a relevant impact on the radiation treatment for localized prostate cancer: treatment gaps do affect the prognosis of oncological diseases, and professionals should moni- risk of biochemical failureBiologisch effektive Dosis und definitiven Strahlentherapie des lokalisierten Prostatakarzinoms: Behandlungslücken tor it during the course of treatment. beeinflussen das Risiko eines biochemischen Versagens. Strahlenther Onkol 2014;190:732–8. [17] Maciá I, Garau M, Solé Monné J, Cambra Serés MJ, Monfà Binefa C, Peraire LM. Compliance to the prescribed overall treatment time (OTT) of curative Conflicts of interest radiotherapy in normal clinical practice and impact on treatment duration of counteracting short interruptions by treating patients on Saturdays. Clin None. Transl Oncol 2009;11:302–11. [18] Khalil AA, Bentzen SM, Bernier J, Saunders MI, Horiot JC, Van Den Bogaert W, et al. Compliance to the prescribed dose and overall treatment time in five randomized clinical trials of altered fractionation in radiotherapy of head and neck carcinomas. Int J Radiat Oncology Biol Phys 2003;55:568–75. Acknowledgements [19] Carruthers S, Pennefather M, Ward L, Giam K, Penniment M. Measuring (and narrowing) the gap: The experience with attendance of Indigenous cancer We thank the radiotherapy technicians involved in the data col- patients for Radiation Therapy in the Northern Territory. J Med Imaging Radiat Oncol 2019;63:510–6. lection for their important role in this project. We are also grateful [20] Hanna TP, Shafiq J, Delaney GP, Vinod SK, Thompson SR, Barton MB. The to the administrative personnel in the different radiotherapy ser- population benefit of radiotherapy: 5-year local control and overall survival vices, and the translation support received from Meggan Harris. benefits. Radiother Oncol 2018;126:191–7.