Oral Cancer patient’s prole and time to treatment initiation in the specialized cancer health care delivery in Rio de Janeiro, Brazil.

Introduction: This paper aims to describe the prole of oral cancer (OC) patients, their risk classication and identify the time between screening and treatment initiation in Rio de Janeiro Municipality. Method: Data were obtained from the healthcare Regulation System (SISREG) regarding the period January 2013 to September 2015. Descriptive, bivariate and multivariate analysis were performed identifying the factors associates with a diagnosis of OC as well as the time to treatment initiation (TTI) differences between groups. Results: From 3,862 individuals with a potential OC lesion, 6.9% had OC diagnosis. OC patients were 62.3 y.o. (mean), 64.7% male%, 36.1% were white and 62.5% of the records received a red/yellow estimated risk classication. Being older, male, white and receiving a high-risk classication was associated with having an OC diagnosis. OC TTI was in average 59.1 days and median of 50 days signicantly higher than non-OC individuals (p=0.007). TTI was higher for individuals older than 60 years old, male, and white individuals and for risk classication red and yellow, nevertheless while in average none of these differences were statistically signicant, the median of individuals classied as low risk was signicantly (p=0.044) lower than those with high risk. Conclusions: Time to treatment initiation (TTI) was higher for OC patients related to non OC. Despite OC conrmed was associated with risk at screening classied as urgent or emergent, a high percentage of OC patients had their risk classied for elective care when specialized care was requested.

because they do not seek health services [2]. These people sometimes have other changes in health, with direct interference in their quality of life and living with their peers [8]. When not treated on time, it is signi cantly mutilating, causing damage to these patients' physical and psychological aspects and directly interfering with the quality of life [9].
However, if diagnosed early, OC has a good prognosis, with the average ve-year survival rate in stages I and II being 77.3%, but 32.2% in stages III and IV [10]. As an early diagnosis of OC is uncommon, with 65 to 85% of cases diagnosed in an advanced stage, the likelihood of cure is reduced [4,[10][11][12].
In Brazil, Primary Health Care (PHC) is supposed to have an active role regarding oral health and its actions include the promotion of oral health and prevention, care, and rehabilitation. In cases of greater complexity, it must be able to act in an articulated and swift manner with the Dental Specialties Center (DSC), which corresponds to secondary care. When appropriate, such as OC diagnosis, rapid treatment initiation in a cancer center is essential. Since 2012, there is in place a federal law (12732/2012) stipulating that in the event of suspected malignant neoplasia, the diagnostic con rmation test must be carried out within 30 days after medical request and speci c treatment must be started within 60 days after the positive cancer diagnosis.
The care trajectory consists of patients' path in the health care network comprising the use of health care resources from the onset of the problem to its outcome [13,14]. It includes making appointments in oral health, the time of return to perform referrals, performing biopsies and obtaining test results, and continuity of care after starting treatment for OC [15], which must act according to a regulation system of health care delivery. The timely care is one of the components of access to the health system. In this study, it was expressed in terms of the time to treatment initiation (TTI).
Rio de Janeiro is a 6.6 million inhabitants municipality located in Southeast region. In 2009 a comprehensive plan for PHC expansion and reorganization was implemented. In addition to a broad set of governmental health facilities comprising all levels of health care, National Institute of Cancer (INCa), the main reference center for cancer care and research in the country, is located in this city.
This study aims to describe the pro le of patients with suspect and con rmed oral cancer, associated factors and time to treatment initiation (TTI) in Rio de Janeiro municipality, Brazil, from January 2013 to September 2015.

Methods
Data were obtained from the healthcare Regulation System (SISREG) regarding the period from January 2013 to September 2015. This study was part from a broader one, a PhD dissertation that had the general aim of evaluating the quality of the care given to patients with oral cancer (OC) in the city of Rio de Janeiro considering the dimensions of access and effectiveness of primary and secondary health care delivery [16].
The database contained all referral records analyzed in the regulation system during the interest period by following medical specialties: general head and neck, head and neck oncology, and dentistry / stomatology.

Variables
The main outcome variable was 'having oral cancer'. The following factors were investigated: age (average and categories), excluding individuals under 15 years old; race (white vs non-white) and rst risk strati cation categorized as high risk (red or yellow) and low risk (green or blue) and TTI. Risk categories are blue -elective care; green -not urgent; yellow -urgent; and red -emergent. The time to treatment initiation was de ned as the time between the rst request date recorded in that database and the last execution date, what indicates the time between the rst appointment and the moment that the individual was seen in his / her de nitive treatment facility (start of OC treatment).
Each encounter referred was recorded in SISREG, thus the same individual can have multiple entries in the system and each record receive a risk classi cation. The overall risk analysis was performed by encounter. To perform the bivariate and multivariate statistical analyzes it was considered the risk evaluation attributed to the patient in the rst encounter.

Analysis
For the rst main outcome, the presence of oral cancer, a descriptive analysis was performed considering the distribution of age, sex and race, as well as bivariate and multivariate analysis. The difference of proportions was expressed by the odds ratio, aiming to identify differences in the demographics and risk determination between cancer and non-cancer patients.
The second outcome was the TTI, and this analysis comprised a descriptive analysis (mean and median) by presence of oral cancer diagnosis. For those with a positive diagnosis of oral cancer, it was performed the non-parametric test of median difference and T-test of the mean to verify if the TTI was different between groups by the age, sex and race. As the distribution of the variable waiting time was heterogeneous, mean and median were very different it was made the option of statistically test the T-test (mean) and the non-parametric tests for median differences.
All statistical analyses were performed at 95% of con dence interval using SPSS V.22.

Ethical issues
The source study was approved by three Research Ethical Committees, the rst from academic institution, the National School of Public Health Sergio Arouca ERC (CAAE 59009416.6.0000.5240), the second from the National Institute of Cancer ERC (CAAE 59009416.6.3001.5274), the hospital where patients were sampled for the face-to-fac interview (data not included in this paper) and the third belonging from the Rio de Janeiro Municipality Health Secretariat ERC (CAAE 59009416.6.3002.5279).

Results
It was found a total of 3,862 individuals with a potential OC lesion in the SISREG. A total of 266 (6.9%) individuals were con rmed as OC. The average age was 52.3 years old for those with no cancer and 62.3 years old in the OC group and there were no cases of oral cancer under 15 years old. Males' percentage was higher among those with OC (64.7%) and the percentage of white individuals was 36.1% and 27.5% among patients with and with no OC respectively (Table 1). From the total 4.764 records, 64.3% were classi ed as blue and 15.1% as red. It was found 27.2% records as red risk among the male while 9.2% among the female. Among those with OC 45.2% were classi ed as red risk, while this percentage was 12.9 among those with no OC. Among those records diagnosed as OC and risk classi ed as red the majority (77.1%) were male ( Table 2). The majority of the patients that received a diagnosis of cancer also received a red risk strati cation.  The average TTI, which is time from the rst request and the appointment for those with no OC was 47.1 days and median of 23 days. Meanwhile, for patients with OC con rmed, the average TTI was 59.1 days and median of 50.5 days until being admitted in a cancer hospital and the difference was signi cant (p = 0.007) ( Table 4). The TTI average was higher for individuals older than 60 years old, male, and white individuals and for risk classi cation red and yellow. It was noticed that the TTI medians also followed the same patters as the means, nevertheless with greater difference. While in average none of these differences were statistically signi cant, the median of individuals classi ed as high risk was signi cantly (p = 0.044) higher than those with low risk.

Discussion
Since 2004, with the expansion of the National Oral Health Policy (PNSB / MS) [3], Brazil has implemented its strategies for prevention, early diagnosis, and control of OC. The PNSB implementation requires easy access to oral health services through the Family Health Strategy (ESF) and the Oral Health Teams as PHC interventions [19].
Articulated actions offered in a timely and resolutive way can prevent the late diagnosis and improve patients' prognostics. In this sense, the OC approach must include the regular screening for early detection; assessment of oral lesions (active search, home visits, speci c campaigns); monitoring of suspected cases; referral services for con rmed cases; and the establishment of partnerships for prevention, diagnosis, treatment, and recovery with universities and other organizations [20].
Public health programs are expected to provide adequate responses to the health problems for which they are intended and evaluated for implementation, access, and outcomes, guiding decision-making [21]. Thus, it is worth asking about the offer of oral health care within the scope of the PNSB in primary, secondary and specialized health services and how patients have covered the services they need.
There is a higher likelihood of receiving a diagnosis of OC for individuals over 60 years old, males, white, which corroborates with results of other studies regarding the pro le of patients and location of lesions. [22,23] When OC diagnosed patients were referred through the SISREG system, the average time to start cancer treatment (average of 50 days and median of 59 days) is within the federal legal time period of 60 days [24]. Nevertheless, the Rio de Janeiro municipality health regulation has a more restrictive TTI rule that establishes that patients diagnosed with cancer must be referred by PHC, via SISREG, for specialized care. When his risk is classi ed as high (red), this requires priority scheduling of up to 30 days. Yellow, scheduling up to 90 days, green and blue, scheduling up to 180 days or more, respectively, was found not to adhere [25]. As well established in the literature, the TTI can be decisive for the progress and incurability of the disease [26,27], and preventable delays should be avoided.
Among the causes for the time elapsed between the rst registration of the patient in the regulation system and the beginning of treatment, one can assume the ow instituted by the health system since the PHC. For cases registered as having the highest risk for OC (red and yellow risk), the time to start treatment was longer than for the others. Serra et al. also found an inadequate organization of referral and counter-referral activities, with many of the patients not being referred by the o cial system, which produces double entry into the sector, resulting in losses or delays in some visits [28]. The opposite, such as decentralization and regionalization of assistance for cancer treatment, facilitate patient access, with an increase in the number of hospitalizations in some locations [29].
Another relevant factor is the di culty in identifying a suspected lesion of these tumors, being most often diagnosed when its size exceeds 2 cm [30]. This situation and other factors can lead to a delay in diagnosis, as indicated by Costa-Jr and Serra [28]: few and nonspeci c symptoms, patients' lack of knowledge about the disease, little familiarity of general practitioners who work in primary care with the diagnosis of cancer and di culty in accessing the health system [31]. The late diagnosis is re ected in the most frequent treatments at the referral hospital, radiation therapy associated with surgery, knowing that the best prognosis is referred to as only surgical treatment [31].
As limitations it is important to mention that the regulation system does not include staging and other clinical information that would be helpful and could support and inform policy changes needed in the regulation and referral. Also, the SISREG is a secondary database and as such, can contain misclassi cations of cases. Finally, only registered individuals that had their treatment through SUS were analyzed.

Conclusions
In this study, it was found that being male, white, and older than 60 increased the likelihood of having a diagnosis of oral cancer. The time to initial treatment was higher for those with con rmed OC diagnosis; nevertheless, TTI is within the time expectation established by Law in the National Health System and there were no disparities regarding age, race and gender. The risk classi cation system seems to not be a factor for speeding up the treatment, furthermore the median TTI of individuals with high risk classi cation was higher and reasons for that should be clari ed in future studies.

Declarations
Ethics approval and consent to participate The source study was approved by three Research Ethical Committees, the rst from academic institution, the National School of Public Health Sergio Arouca ERC (CAAE 59009416.6.0000.5240), the second from the National Institute of Cancer ERC (CAAE 59009416.6.3001.5274), the hospital where patients were sampled for the face-to-fac interview (data not included in this paper) and the third belonging from the Rio de Janeiro Municipality Health Secretariat ERC (CAAE 59009416.6.3002.5279). Nevertheless, we use secondary data for this current manuscript and an informed consent is not applicable.

Consent for publication
Not applicable Availability of data and materials The datasets generated and/or analysed during the current study is not publicly available due [REASON WHY DATA ARE NOT PUBLIC] but are available from the corresponding author on reasonable request.

Competing interests
Authors have no competing interests

Funding
This study had no funding.
Authors' contributions MGD, VLL and ACF are the guarantors of the paper and takes responsibility for the integrity of the work as a whole. MGDC was responsible for obtention of the dataset. All authors were responsible for the study design, data analysis, interpretation results, writing, and revisions and approved the submitted version.