Introduction

Growing cancer morbidity and mortality are overburdening healthcare systems in low- and middle-income countries (LMICs) [1]. In 2018, nearly 60% of new cancer cases and 70% of cancer-related deaths occurred in adults ≥ 20 years of age in LMICs [2]. Physical and psychological complications and treatment side effects negatively affect day-to-day functioning, social life, and well-being of cancer patients, requiring supportive care to address their needs and ensure optimal experiences and outcomes [3, 4]. However, most patients do not receive proper supportive care (SC) to cope with their physical, psychosocial, and information needs [3].

Additionally, LMICs have wide inequalities in cancer incidence and mortality, with shorter survival rates and lower quality of life among socially disadvantaged cancer patients in contrast with those who are better off [5]. Social inequalities are systematic disparities in the health status of groups with underlying social disadvantages. Social inequalities in health behaviors (e.g., smoking) and in cancer mortality are widely recognized. In contrast, the evidence on social inequalities in SC-needs and quality of cancer care in LMICs is scarce [5,6,7].

Although social inequalities arise outside the health system, healthcare is a significant determinant of health able to bridge social disparities by providing access to affordable, patient-centered, and high-quality care sensitive to the social needs of vulnerable groups [5, 6]. Patient-centered care (PCC) aims at meeting needs, expectations, and preferences of patients through respectful, continuous, and coordinated care [8, 9]. Moreover, PCC improves satisfaction and quality of life, reduces healthcare expenditures, and can reduce the supportive care needs of patients [6, 10, 11].

Mexico is an upper-middle-income country with the highest rate of social inequality across OECD countries [12]. In 2017, 30% of people aged ≥ 15 years had not completed elementary school. The widest educational lag is visible in the poor [13]. Besides, the fragmented Mexican healthcare system is inequitable and separates the population according to their employment status without considering their health needs. Formal labor market workers and their families (nearly half of the Mexican population—65 million people) [14] have access to social health insurance through the social security institutes; most are affiliated with the Mexican Institute of Social Security (IMSS). The underprivileged population, informal workers, and the unemployed receive care at the Ministry of Health (MoH) facilities. After the end of Seguro Popular in December 2019, this group lacks health insurance. Disparities are also visible in the health benefits. IMSS affiliates have access to a broader package of healthcare services than people receiving care at the MoH.

In 2018, Mexico reported 183,000 new cancer cases and 81,000 cancer deaths in persons ≥ 20 years of age [2]. Breast, prostate, colorectal, and hematologic cancers were the most common causes of cancer-related morbidity and mortality [2]. Studies in breast and other solid cancers have shown a high proportion of IMSS and MOH patients with unmet SC-needs of health system information and physical and psychological care [15, 16]. However, social inequalities in SC-needs and PCC-quality that cancer patients receive at MoH and IMSS facilities have not been ascertained.

This study had two objectives: (1) to evaluate the educational-related inequalities in SC-needs and PCC-quality for cancer patients with and without social health insurance in Mexico and (2) to assess the association between the PCC-quality and SC-needs.

Methods

We conducted a cross-sectional survey from April 2018 to September 2019 in two of the largest oncology hospitals in Mexico City selected by convenience sampling., one belonging to the MoH and the other to the IMSS.

The study population comprised outpatient cancer patients aged ≥ 18 years with breast, colorectal, prostate, and hematologic cancers (leukemia, lymphoma, or multiple myeloma). We included patients with at least one hospitalization during the last year, ≤ 5 years since diagnosis, and without mental impairment. Four fieldwork-trained nurses interviewed the patients after their medical consultations, after they had met the inclusion criteria, agreed to participate, and signed informed consent forms. Two field coordinators verified patients’ diagnosis and treatment in their health records.

Study variables

The study has one independent variable (educational attainment) and two dependent variables (SC-needs and PCC-quality).

We defined educational attainment as the years of completed schooling categorized as completed elementary school or less, secondary school, and high school or higher.

To measure SC-needs, we used the short-form Supportive Care Needs questionnaire (SCNS-SF34) [17] validated in Mexico [15], which consists of 33 items grouped into five domains: (1) psychological needs, (2) health system information needs, (3) physical and daily living needs, (4) sensitivity of healthcare professionals to patient care needs, and (5) sexuality-related SC-needs. Each item has a 5-point Likert response option scale where 1 corresponds to the absence of need and 5 to high need for support. Scores for each domain were calculated according to McElduff et al. recommendations, with standardized scores ranging from 0 (no need) to 100 (high need) [18]. The Mexican SCNS-SF has a five-domain structure that accounts for 59% of total variance; the internal consistency of the factors ranges from 0.78 to 0.90; additionally, the scale has good convergent and discriminative validity [15].

To measure PCC-quality we used a patient-centered quality of cancer care questionnaire (PCQCCQ) [19] validated in Mexico [20]. This questionnaire has 30 items and encompasses five domains: (1) timely care, (2) clarity of information, (3) information for treatment decision-making, (4) care to address biopsychosocial needs, and (5) respectful and coordinated care. The score for each domain was calculated by reversing the response options, adding all subscale items and dividing them by the number of items in each factor with a minimum score of 1 and maximum of 4 per domain. The five domains of the Mexican PCQCCQ account for 96% of total variance; the internal consistency of the scale is 0.90, with Cronbach’s alpha values ranging from 0.73 to 0.90 among the domains; the scale can differentiate by “known groups” (e.g., age) [20].

Study covariates comprised patient sociodemographic characteristics (gender, age, and marital status), clinical history (time since diagnosis, cancer type and stage, treatment in the last month, and anxiety, depression, and chronic illnesses other than cancer).

We categorized patient age (≤ 45, 46 to 64, and ≥ 65 years), cancer type (breast, colorectal, prostate, and hematological), cancer stage/or risk (early stage (I–II) for solid cancers or low and standard risk for hematological cancers, advanced stage (III–IV) for solid cancers or high and very high risk for hematological cancers), and time since diagnosis (≤ 6 months, 7–12 months, >1 to 5 years).

We measured anxiety and depression using the Hospital Anxiety and Depression Scale (HAD) [21] composed of 14 items previously validated in Mexico with cancer patients [22]. Each item has a 4-point Likert scale response that ranges from 0 to 3. The HAD has two factor structure that explains 48% of the variance, with the internal consistency of 0.79 and 0.80 for each subscale [22]. A cut-off value of ≥ 11 points in each domain was chosen to screen for anxiety and depression as indicated by the authors of the original scale [21].

Sample size and statistical analysis

We included a minimum of 10 participants per covariate in the multiple regression analysis [23].

We performed descriptive and exploratory analyses and found that the dependent variables (SC-needs and PCC-quality domains) did not have a normal distribution. We compared the medians of SC-needs and PCC-quality between MoH and IMSS using the Kruskal–Wallis test.

To determine the association between independent and dependent variables, we dichotomized the variables as high and low SC-needs and PCC-quality by using the 75th percentile of the total sample as a cut-off value. The distribution of variables supported this decision, including the low frequency of patients at 85th, 90th, and 95th percentiles.

First, we created five high SC-needs variables: (1) psychological needs ≥ 60.0 points, (2) health system information needs ≥ 68.2 points, (3) physical and daily living needs ≥ 65.0 points, (4) patient care needs ≥ 35.0 points, and (5) sexuality-related needs ≥ 50.0 points. Then, we built five high PCC-quality variables: (1) timely care = 4.0 points, (2) clarity of information = 4.0 points, (3) information for treatment decision-making ≥ 3.4 points, (4) care for biopsychosocial needs ≥ 2.2 points, and (5) respectful and coordinated care ≥ 3.8 points. Second, we compared participants’ characteristics between MoH and IMSS and between those with and without high SC-needs and those with and without high PCC-quality using the chi-square test. Third, we estimated the crude odds ratio (OR) and 95% confidence intervals (95%CI) between the independent variables, covariates, and each dependent variable. Fourth, to determine the association between the independent (educational attainment) and dependent variables (high SC-needs and high PCC-quality domains), we built multiple logistic regression models with simultaneous inclusion of independent variable and all conceptually and clinically relevant covariates, as recommended by VanderWeele [24]. This step allowed controlling for the covariates to provide an unbiased estimate of the educational inequalities.

We chose the covariates through literature review. Previous studies have found that gender (being a woman), age (being younger), specific cancer types, advanced disease stage, short time since diagnosis, and anxiety, depression, and other comorbidities are associated with high SC-needs [3, 16, 25,26,27,28,29]. PCC-quality also affects SC-needs [6]. Therefore, we controlled the analysis of SC-needs including the above-mentioned covariates.

In the case of PCC-quality models, prior research reported that sociodemographic and clinical characteristics such as younger age, residence in poor areas, time since diagnosis, comorbidity, type of cancer, and type of health services were associated with cancer care quality [30]. We used the software Stata 14.0 (Stata Corp, College Station, TX, USA) for the analysis; p < 0.05 was set as statistically significant.

Ethics approval

The study was approved by the IMSS National Research and Ethics Committee (registry number R-2017-785-042).

Results

In both institutions a high proportion of eligible cancer patients agreed to participate (IMSS 87.9%; MoH 92.4%). The main reasons for declining to participate were lack of interest in answering the survey, lack of time, and fatigue, weakness, or pain. There were no statistically significant differences between those who accepted or declined to participate in the study regarding sex, age and cancer diagnosis.

More women participated at MoH (65.2%) compared with IMSS (49.2%). Half of IMSS participants had completed high school compared with 25% at MoH, where 74.9% completed only secondary school or less. The average age of participants was higher at IMSS than at MoH (57.8 vs. 50.2 years). Most participants were married (69.9% IMSS; 59.4% MoH) (Table 1).

Table 1 Characteristics of the study population
Full size table

At IMSS, 31% of participants had hematologic cancers, 23.4% breast or colorectal cancer, and 22.2% prostate cancer. At MoH, 40.8% of participants had breast cancer, 30.9% hematologic cancers, 16.3% colorectal cancer, and 12% prostate cancer. Most participants had been diagnosed with cancer ≤ 6 months prior (IMSS 40.6%; MoH 43.4%) or between 7 and 12 months prior (IMSS 45.2%; MoH 45.4%). In both institutions, most patients with solid cancers were at advanced stages (III–IV) or at high/very high risk in the case of hematologic cancers (67.9% IMSS; 63.4% MoH). More IMSS patients had undergone surgery in the last month (20.5% IMSS; 7.4% MoH), while more MoH patients had been receiving chemotherapy or radiotherapy (61.4% IMSS; 73.9% MoH). Comorbidities were more common in IMSS participants (48.4%) than in MoH participants (28.9%). Less than 20% of patients in both institutions had anxiety (16.5% IMSS; 13.9% MoH) or depression (14% IMSS; 10.9% MoH) (Table 1).

SC-needs were higher among MoH patients. On a 100-point scale, the highest needs were health system and information related (median: MoH 54.5 points; IMSS 34.1 points), followed by physical and daily living needs (MoH and IMSS 40.0 points) and psychological needs (35.0 points MoH; 32.5 points IMSS). Based on the 75th percentile, there were more patients with high health system and information and care SC-needs at MoH compared with IMSS (33.7% MoH; 16.8% IMSS and 34.7% MoH; 22% IMSS, respectively).

The PCC-quality was higher at IMSS than at MoH. On a 4-point scale, the respectful and coordinated care domain got the highest quality score (3.5 points IMSS; 3.3 points MoH), and the lowest was for care to address biopsychosocial needs (1.6 points IMSS; 1.4 points MoH). Based on the 75th percentile, more IMSS patients perceived high PCC-quality related to the clarity of information (IMSS 42.2%, MoH 29.4%), and fewer patients experienced high PCC-quality that addressed their biopsychosocial needs (IMSS 24.1%; MoH 18.5%) (Table 1).

High SC-needs varied significantly by schooling, gender, age, and other participant characteristics (Table 2). Compared with men, women at IMSS reported psychological and physical SC-needs more often, while men reiterated high sexuality SC-needs. In both institutions, participants with elementary schooling had more physical and daily living SC-needs than their counterparts. IMSS participants with secondary education reported psychological SC-needs more often, and those with high school reported sexuality SC-needs. Younger IMSS participants reported higher psychological, health system, and care needs compared with their counterparts. In both institutions, younger and married participants expressed high sexuality SC-needs.

Table 2 Cancer patients with high supportive care needs by participant sociodemographic and clinical characteristics
Full size table

As for the clinical characteristics, IMSS participants with colorectal cancer reported psychological needs more often, while patients with hematologic malignancy reported more health system and physical SC-needs. In both institutions, participants with prostate cancer expressed high sexuality SC-needs. Also, IMSS participants with ≤ 6 months since diagnosis and those who underwent surgery reported higher physical SC-needs than their counterparts. MoH patients at early cancer stages reported health system SC-needs more often.

At both institutions, patients with anxiety and depression reported high SC-needs more often than their counterparts in all domains. Patients with chronic comorbidities at IMSS reported higher physical SC-needs but fewer sexuality SC-needs. Generally, patients with high PCC-quality reported fewer SC-needs.

Table 3 shows patient perception of PCC-quality according to their characteristics. Compared with women, men at MoH reported timely care less often. In both institutions, participants with elementary schooling reported less timely care, clarity of information, and care that addressed their biopsychosocial needs compared with those with higher education. Furthermore, MoH participants with elementary schooling reported receiving less information for treatment decision-making, and married patients experienced less high-quality care that addressed their biopsychosocial needs.

Table 3 High quality of patient-centered health care by participant sociodemographic and clinical characteristics
Full size table

Compared with other cancers, IMSS participants with colorectal cancer and MoH participants with prostate cancer reported timely care less frequently, while in both institutions, prostate cancer patients experienced high-quality care that addressed their biopsychosocial needs less often. Moreover, IMSS participants with ≤ 6 months since diagnosis reported less clarity of information and care that addressed their biopsychosocial needs, while in both institutions, these participants received less information for treatment decision-making. MoH participants at early cancer stages (III–IV) reported clarity of information and care that addressed their biopsychosocial needs less frequently. Moreover, those with surgery in the last month experienced timely care more frequently, while those with chemotherapy did not.

At IMSS, patients with anxiety and depression perceived high PCC-quality less often in almost all domains. At MoH, there were no statistically significant differences in high PCC-quality between patients with and without depression, but patients with anxiety reported timely care, respectful and coordinated care, and attention to their biopsychosocial needs less often. There were no differences in PCC-quality reported by patients with and without other comorbidities.

The results of the multivariate analyses assessing association between education and high PCC-quality with SC-needs (Table 4) show that compared with cancer patients with a high school education, elementary school patients at MoH had increased probability of psychological SC-needs (aOR1.95, 95%CI 1.09; 3.47), whereas patients with both elementary and secondary school had high health system and information SC-needs (aOR 1.87, 95%CI 1.09; 3.19; and aOR 1.92, 95%CI 1.14; 3.23, respectively). IMSS elementary school participants had low probability of sexuality SC-needs (aOR 0.41, 95%CI 0.27; 0.62). In both institutions, perceived clarity of information, respectful and coordinated care, and care for biopsychosocial needs reduced the likelihood of high SC-needs. These associations were independent of other sociodemographic and clinical covariates.

Table 4 Association of education and high quality of patient-centered care with supportive care needs of cancer patients
Full size table

We present the information on the magnitude of the association of covariates with SC-needs in Electronic Supplementary Material 1. Briefly, compared with men, women at IMSS had a low probability of high sexuality SC-needs, while women at MoH had elevated probability of psychological SC-needs. Being single or divorced decreased the likelihood of sexuality SC-needs for both IMSS and MoH patients. Moreover, compared with participants ≥ 65 years of age, younger IMSS patients had a higher probability of health system, care, and sexuality SC-needs. Clinical covariates significantly associated with high SC-needs were cancer type, time since diagnosis, cancer stage, surgery, comorbidity, anxiety, and depression. For instance, hematologic cancer was associated with an elevated probability of physical SC-needs among IMSS patients. Hematologic and colorectal malignancies were associated with decreased sexuality SC-needs at IMSS. At MoH, having breast cancer was associated with lower sexuality SC-needs. Shorter time since cancer diagnosis was associated with decreased physical SC-needs in MoH patients. Being at an advanced cancer stage was associated with a high probability of physical SC-needs in IMSS patients and reduced probability of health system SC-needs in MoH patients. Having surgery was associated with higher physical SC-needs among IMSS patients. In both institutions, anxiety and depression were associated with an increased probability of high SC-needs in most domains.

Table 5 depicts the results of the multivariate analyses evaluating education-related inequalities in high PCC-quality independent of other covariates. Having elementary schooling or less was associated with a decreased probability of timely care at IMSS (aOR 0.70, 95%CI 0.49; 0.99), clarity of information received in both institutions (aOR 0.50, 95%CI 0.36; 0.69 at IMSS and aOR 0.53, 95%CI 0.32; 0.86 at MoH), quality of information for treatment decision-making at MoH (aOR 0.53, 95%CI 0.32; 0.86), and care for biopsychosocial needs at MoH in patients with elementary (aOR 0.44, 95%CI 0.25; 0.78) and secondary schooling (aOR 0.51, 95%CI 0.30; 0.86).

Table 5 Education-related inequalities in quality of patient-centered cancer care
Full size table

Electronic Supplementary Material 2 displays the information on the magnitude of the association of covariates with PCC-quality. There were no differences in high PCC-quality due to other sociodemographic covariates. In both institutions, having breast cancer was associated with a higher probability of timely care; moreover, at IMSS it was associated with care for biopsychosocial needs and at MoH, with respectful and coordinated care. However, recent diagnosis (≤ 6 months), anxiety, and depression were statistically significant clinical covariates associated with decreased probability of high PCC-quality in most quality of care domains.

Discussion

This study contributes to gaining an in-depth understanding of the inequalities of cancer care in Mexico. First, it reveals the magnitude of existing educational and health insurance-related disparities in SC-needs and PCC-quality. Second, it identifies that high PCC-quality reduces the likelihood of high SC-needs. This positive effect happens when providers address biopsychosocial needs, deliver precise information, and provide respectful and coordinated care.

Mexico has a highly fragmented healthcare sector that ties the right to healthcare to employment status, thus having separate providers for people working in the formal labor market that are affiliated with social security institutes, such as IMSS, and for those underprivileged and working in the informal labor market that lack health insurance and are cared for by the MoH. Over the last two decades, Mexico had made important progress towards universal health coverage for people without social security; however, recent changes in Mexican health policies, such as the end of Seguro Popular, create uncertainty regarding the goal of reducing inequities. This study signals that healthcare-related inequalities persist. Our study found that cancer patients without social security treated at a MoH oncology hospital perceived high SC-needs and low PCC-quality more often than patients with social security health insurance treated at IMSS. This finding is similar to a study from the USA that found insurance related disparities in SC-needs among lung cancer patients, due to the fragmentation of the health sector [31].

PCC-quality is a challenge for oncology healthcare services within the resource constraints of the Mexican health system, where the model of care and services is designed without the necessary personnel and processes to satisfy SC-needs properly. Previous studies comparing the quality at IMSS and MoH that reported inconsistent results signal the necessity to improve governance of quality of care among these institutions. For instance, one study found better quality of reproductive services for adolescents at social security institutes than at the MoH [32], while another study reported higher quality diabetes care at MoH facilities compared with social security institutes [33].

Social inequalities are beyond the control of the health system; however, properly designed comprehensive and high-quality oncology services are helpful to mitigate such inequities. In our study, lower levels of education were associated with increased risk of high psychological and health system and information SC-needs and with a low likelihood of high PCC-quality. Several studies from high-income countries (HICs) have found that the educational level of cancer patients do not have much impact on their experiences with patient-centered care [34, 35]. Meanwhile, high PCC-quality in our study was associated with a decrease in SC-needs, especially when providers address patients’ biopsychosocial needs, communicate information clearly, and provide respectful and coordinated care. These results signal the relevance of focusing interventions on improving PCC-quality and decreasing SC-needs in cancer patients with low education in Mexico.

Cancer patients with low education have shorter survival and poorer quality of life than those with higher schooling [5, 7, 36, 37]. Moreover, studies in HICs have identified pro-rich and pro-educated inequities in specialist care utilization [38, 39], clinical process quality, and patient experiences with public healthcare services [40]. These findings signal the interconnections among patient education, PCC-quality, and SC-needs that might explain education-related survival inequalities. Consequently, high PCC-quality that addresses SC-needs is an essential attribute of comprehensive cancer care [41].

Patients with low education are less prepared to interpret health information—this negatively affects their ability to navigate the health system and adhere to treatment recommendations [42]. Therefore, the health system and health providers must be prepared to effectively communicate with people with low education and inadequate health literacy [42]. Clarity of information is central to effective patient-provider communication. Poor communication negatively affects patient trust, satisfaction with care [41], pain control [43], commitment to treatment decisions [44], and caregiver burden [45]. Providing thorough and jargon-free explanations, verifying understanding, and encouraging patients to express their concerns reduce anxiety and enhance patient–provider communication [46]. Moreover, patient-centered cancer care should comprise an integrated approach to fully address the complexity of cancer [47].

The study has several strengths and limitations. First, this was a cross-sectional study that does not allow for making causal inferences or identifying the direction of association between study variables. Second, the sample included patients from two oncology hospitals in the country, limiting its generalizability. The study’s strengths include having a large sample of cancer patients from two major healthcare providers in Mexico, using validated questionnaires, and addressing education and health insurance-related inequalities in SC-needs and PCC-quality for cancer patients—a poorly assessed issue in LMICs.

Conclusion

Cancer patients with low education and without social security-based health insurance have high supportive care needs and receive substandard quality of patient-centered care. Health services should face these challenges to reduce inequalities in Mexico.