Prevalence of multimorbidity in adults with cancer, and associated health service utilization in Ontario, Canada: a population-based retrospective cohort study

Multimorbidity (MMB), defined as the co-occurrence of multiple chronic conditions, is a public health crisis, challenging healthcare systems and affected individuals and their families [1‐5]. The risk of MMB is particularly high among people with cancer [4, 6‐10]. In Ontario, Canada, more than 75% of people with cancer have at least one of 16 other prevalent chronic conditions [4]. Rates of MMB among cancer patients, however, vary by cancer type [6, 11, 12]. Similarly, the chronic condition(s) that co-occur with cancer also vary by cancer type and can have different impacts on patient outcomes, including death [11, 13, 14].

MMB impacts people with cancer directly through an increased physiological burden of disease and/or indirectly through healthcare factors related to treatment decision-making and the increased complexity of care [15]. For example, MMB can complicate and delay cancer and other diagnoses, increase treatment complexity, risk of polypharmacy, and/or create other redundancies in care [16‐20]. Cancer patients with MMB are less likely to be offered active or curative therapies [21‐23], despite growing evidence that such treatments can be tolerated and effective for some patients [24‐26]. People with cancer and MMB likely also experience increased burden from managing multiple appointments and care regimens, which may result in confusion, stress, difficulty adhering to treatment protocols, and poor quality of life.

Another critical challenge to effectively treating and improving outcomes for people with cancer and other chronic conditions is that most care plans follow a single disease-oriented approach [10, 18, 27, 28]. In such single-disease models, the management of co-occurring conditions often takes a backseat to cancer even when coordinated management of chronic conditions could help improve patient outcomes and quality of life [29, 30]. Well-documented gaps in clinical guidelines and care management frameworks demonstrate the need for intentional integration and coordination of cancer and chronic disease care, including clear direction to providers who may lack clarity about their roles and responsibilities [31‐33] for patients receiving services spanning siloed specialized and primary care settings [2, 34‐36].

In order to improve care and outcomes for people with cancer and MMB, health care systems must recognize and reorganize care delivery in order to provide whole-person care. However, few large epidemiological studies have examined the overall extent of MMB among people with cancer of all types to provide guidance about how care should be improved. Similarly, researchers have not broadly documented the potential impact of cancer MMB on health service utilization at the general population level in order to motivate care delivery reform.

In this paper, we address these research gaps by describing the co-occurrence of cancer of any type with 16 common chronic diseases, as well as the level of MMB associated with patients’ health service utilization in an Ontario, Canada population-based, retrospective cohort study. In a second, companion paper using this same data set, we examine which conditions are most likely to co-occur with cancer and how specific disease clusters impact patient outcomes (Koné et al., under review).

Our study population included 601,331 adults diagnosed with cancer between 2003 and 2013; 49% were females and 55.7% aged 65 years or older. The most common types of cancer identified in the sample were (in descending order): Prostate, breast, colon/rectum, and lung/bronchus (Table 1).

This population-level study is among the first to examine the co-occurrence of cancer of all types and 16 other chronic conditions managed outside of the cancer care system. While research on MMB continues to grow, there are only few studies assessing the number of co-occurring conditions among people with cancer, and those studies have typically included only a few specific cancer sites [7, 8, 11]. By describing the extent of MMB within cancer of all types, we aim to inform Canadian and global efforts to improve quality, efficiency, and patient outcomes and experiences in complex (i.e., multimorbid) cancer care.

Overall, the prevalence of MMB in cancer patients is among the highest [4, 41]. Similar to previous studies [7, 11], we found that cancer MMB was extremely common (91%) in our population of patients diagnosed with cancer between 2003 and 2013. MMB was also present across all stages of cancer, with some indication that MMB prior to cancer was related to diagnosis at later stages, but the overall proportion of extensive MMB was lower at later cancer stage, likely due to survival bias or lack of focus on other conditions following cancer diagnosis. Our data also showed that rates of cancer MMB remained stable over time, indicating the ongoing need to address MMB effectively.

While patients are likely to exhibit multiple conditions before cancer, MMB continues to increase substantially after a cancer diagnosis. In our study, we found that nearly half of all people with cancer were diagnosed with an additional chronic condition during follow-up. This is consistent with data from Leach and colleagues [7] in which cancer patients reported the emergence of approximately 1.9 new conditions after diagnosis of one of five common cancer types.

Our data is also consistent with previous studies showing no sex difference in MMB rates but increased rates and levels of MMB among older age groups [9, 11]. Notable from our study, however, are findings that cancer MMB was observed across all age groups, including high rates of cancer MMB  with two or more conditions in young adults (ages 18–44 years). This suggests that MMB is not only an issue for older adults but a growing issue for adults of all ages [4, 41‐43].

Cancer sites with the highest prevalence of MMB were myeloma, prostate, urinary system, lung/bronchus and Leukemia. For example, those with four or more conditions beside cancer represented a substantial proportion for all sites of cancer from 19% (cervical cancer) to 53% (myeloma). Similar to our study, Fowler et al. [11] found that MMB was higher among people with lung cancer compared to the other four cancers included in their study, with 67% of lung cancer patients having one or more comorbidities. Different from Fowler and colleagues, however, is that our overall rates of MMB are higher. An important difference between our methods is that Fowler et al. [11] considered four possible co-occurring conditions instead of the 16 included here. Other contextual factors, such as public health screening initiatives, may also be important for between-population and between-cancer rates of MMB. For example, in Ontario, Canada, cervical cancer was among the sites with the lowest prevalence of MMB in our study (66% before cancer and 82% overall) but it is also the target of a province-wide screening initiative for women as young as age 21 [44] who are less likely than older adults to have MMB.

We also analyzed the impact of MMB on multiple aspects of the HSU. While adequate access to primary care may contribute to better care management and patient outcomes, high use of ED and hospital services, and ultimately death, reflect poor patient health which may be aggravated by the presence of multiple conditions. Not surprisingly, both mortality and HSU were positively associated with increasing MMB levels in our study, regardless of patients’ age, sex, cancer type or cancer stage. This is in line with previous research showing that higher MMB was associated with a higher risk of death or lower survival among patients with specific cancer types [6, 12, 45]. Legler et al. (2011) also found a positive association between high patient Charlson Comorbidity Index and increased admission to ED, hospital, and ICU [46]. Unlike prior work, a unique contribution of this paper is that we examined the relationship between MMB level and HSU. In those analyses, we observed that increasing MMB level had a greater negative impact on HSU among younger people: the difference between cancer only and the highest level of MMB was on average 1.8 ED visits and 7.1 PC visits in young adults compared to 0.4 and 4.3 among people 65 years and older. The trend was opposite for mortality: increasing MMB was not as strongly linked to risk of death among younger adults. More research examining potential contributing factors, such as early screening and identification through regular health service use, for example, is needed to better understand cancer MMB in younger adults.

Lastly, our data showed that individual conditions had varied impacts on patients’ health outcomes, depending on the level of MMB. Conditions frequently identified as co-occurring with cancer in previous studies include hypertension, COPD, diabetes, CVD, and CHF [11, 46]. These conditions were also among the most frequent in our study population. The most common co-occurring conditions (hypertension, arthritis, anxiety, diabetes) often had the greatest impact on outcomes; however, other less prevalent conditions are also worth considering. Overall, the relationship between cancer, mental health (including substance use disorder) diagnoses, and HSU is in need of further study. Our data showed increased PC encounters among people with co-occurring anxiety disorder, which was also associated with fewer hospitalizations and a lower risk of death. While it well known that psychiatric conditions are a significant driver of HSU (e.g., [47]), conditions like anxiety that increase preventive contacts with the healthcare system may represent a protective form of MMB. Other conditions such as dementia or psychosis that limit patient capacity for self-care, in comparison, may increase higher-acuity service use, such as ED visits and hospitalizations, signaling the need for more intensive preventive care and/or illness self-management supports. Overall, these findings suggest that while MMB is important to understand overall, unique combinations of co-occurring conditions are likely to have differential effects on HSU and patient outcomes, and thus require further study.

People with cancer are very likely to experience MMB over all and at very high levels, regardless of age and cancer type – including young adults whose risk for MMB may have been previously overlooked. The findings in this report can inform health system advances towards person-centered care as it describes the overall nature MMB for the whole range of cancer types. At the same time, our findings also show that there is wide variation in the impact of individual conditions within and between MMB levels, suggesting that it is crucial to assess the role of each condition within a MMB lens. This could, for example, include routine assessment of chronic conditions that commonly co-occur with cancer and coordinated care pathways that reduce treatment burden and increase access to targeted supports (e.g., mental health; illness self-management) for patients likely to experience the need. More generally, however, our findings show that at higher levels of MMB, individual conditions have less of an impact on outcomes that MMB itself, again highlighting the need to improve how care for people with MMB is envisioned and operationalized. In the next paper, we describe the relationship between specific, observed clusters of cancer and chronic diseases and their relationship to HSU and mortality risk to help guide next steps in advancing MMB care.