Background

According to the WHO, global cancer incidence is currently 19 million and is expected to increase to an estimated 30 million in 2040 [1, 2]. The rise in new cancer cases is met by improved survival rates, primarily due to optimized treatment options and early disease detection. These improved survival rates are often accompanied by chronicity of the disease and thus a reduced quality of life (QoL) due to psychological impairment, physical dysfunction, chronic pain, sleep disturbances, and fatigue [3].

Complementary and alternative medical care options have gained much attention in recent years and for the majority of cancer patients, they have become an important component of their medical care [4]. This refers to treatment methods that are understood as complementary to scientifically based methods of medicine and are therefore subsumed under the term complementary medicine. The term CAM (complementary and alternative medicine) encompasses a group of different medical and health-promoting methods, practices and products that are generally not part of conventional medicine [5]. A recent review of 12 studies showed that 25%– 80% of all cancer patients use CAM services as part of their cancer treatment [6].

Utilization depends on sociodemographic factors and medical characteristics. Among breast cancer (BC) patients, relatively high levels of CAM use are found among younger married women with higher levels of education and income, as well as among patients at an advanced disease stage or relapse and those undergoing chemotherapy, radiotherapy, or hormone therapy [7–10]. In another study of 288 BC patients in Korea, it was shown that CAM was more likely to be used by patients who reported a higher symptom burden compared to patients with a lower symptom burden [11].

For gynecological cancer (GC) patients, there are significantly fewer studies, and the results are inconsistent. In a review by Akpunar et al. that included data from 1,978 patients and summarized results obtained in studies from developed countries, there were statistically significant correlations between CAM use and younger age and higher education levels [12]. A study from Turkey with a small sample size of 67 patients [13], as well as a larger study from Thailand [14] with 370 study participants did not find differences in sociodemographic or medical factors. A number of studies [7, 11–17] substantiate the following motives for CAM use in BC and GC patients: an intent to "strengthen the immune system," "reduce cancer therapy-related side effects," "increase physical and mental fitness," and thereby "improve their quality of life", as well as "cure their cancer." Further central motives were: "regaining control over the disease" and the desire to "actively participate in the treatment process" [18].

CAM offerings are broad. According to a classification by the NCCIH (National Center for Complementary and Integrative Health), they include:

1. natural products such as herbs, vitamins, minerals, and supplements,
2. mind-body practices such as yoga, meditation, and acupuncture, and
3. complementary approaches such as Traditional Chinese Medicine (TCM), Ayurveda, and homeopathy [5].

Worldwide, the most popular CAMs among cancer patients in general are natural products of various plant and processed types, as well as vitamins and minerals [6]. The products used depend strongly on the region. In Arab countries, for example, the preferred CAMs tend to be religious in nature, such as prayers and recitation of the Quran [15, 19]. In European countries, vitamins and minerals are most common in BC patients, as well as relaxation methods, homeopathy, and mistletoe therapy [7, 9]. In GC patients, herbal medicine, vitamins and minerals are favored [12]. Several CAMs are often used together or successively [8, 10, 13, 16]. With regard to satisfaction with the CAM offers, a high mean satisfaction (60%) was shown in BC and GC patients [8, 20].

Material and methods

Participants and procedures

In this sub-sample analysis, we use data from a large, multicenter, epidemiological, cross-sectional study, the methods of which have been described in more detail elsewhere [23, 24]. Patients were recruited while receiving treatment from oncological inpatient clinics at acute care hospitals, specialized outpatient cancer care facilities, and cancer rehabilitation centers across five study centers in Germany (Freiburg, Hamburg, Heidelberg, Leipzig, and Würzburg). Inclusion criteria consisted of being a cancer patient, between the ages of 18–75, with a malignant tumor. Patients across all tumor entities and disease stages were included. All participants provided written informed consent before taking part in the study. Exclusion criteria were: age younger than 18 or older than 75 years, severe cognitive or physical impairment, language barriers. The study complied with the Declaration of Helsinki and was approved by the research ethic committees of the local medical association in each study center.

Of 5,889 eligible patients, 4,020 (68%) agreed to participate in the study (Fig 1). The primary reasons given for non-participation were lack of interest (55%) and symptom burden (33%). In this analysis, we considered only female patients suffering from BC (897 patients) or GC (317 patients, in particular: 79 cervical, 70 endometrial, 131 ovary and fallopian tube, 37 vulva and vagina).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of the sample.

https://doi.org/10.1371/journal.pone.0285718.g001

Results

Sample characteristics

Fig 1 shows the study enrollment. A subsample of 1,214 female patients were selected, of whom 897 suffered from BC and 317 from GC. Sociodemographic and medical characteristics of BC and GC patients are presented in Table 1, column A. In summary, there were no differences regarding age, vocational education, and monthly income between the cancer groups. By contrast, significant differences were noticed in relationship status (BC 78.3% vs GC 72.3%, p<0.05), occupational situation (BC 49.6% employed, GC 40.8% employed, p<0.003), time since diagnosis (BC 18.2 month vs. GC 11.2 month, p = 0.001), disease stage, past/current treatment, setting, and the distress thermometer (Table 1, column A: sample characteristics).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1.

Sample characteristics (A) and comparison CAM user vs. non-CAM user (B).

https://doi.org/10.1371/journal.pone.0285718.t001

Frequency of CAM use in BC and GC patients and perceived benefit of CAM use

Fig 1 shows that, overall, 46.5% of the patients used CAM. As Fig 2 shows, 21.1% of BC patients and 20.2% of GC patients used exactly one type of CAM, and about 1 in 10 women used two types (BC 12.6%, GC 9.8%). 60% of respondents in both disease groups rated the methods as quite a bit or very helpful, and a small proportion of 13% rated them as a little or not helpful. A comparison of the groups indicates that BC patients were significantly more likely to use CAM than patients with GC (48.6% vs 40.7%, p = 0.015). The proportion of non-users of both groups together is 53.5%.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Prevalence of CAM use in BC and GC patients and its benefit.

https://doi.org/10.1371/journal.pone.0285718.g002

Predominantly used types of CAM services in BC and GC patients

Fig 3 shows the most frequently used types of CAM in both groups: vitamins/minerals/micronutrients, homeopathy, and mistletoe therapy. In contrast to that, Bach-flowers therapy, kinesiology, and neural therapy were hardly used. There was a statistically significant difference in the use of vitamins/minerals/micronutrients between BC patients and GC patients (BC 41.2% vs GC 30.0%, p<0.001) and in the use of enzymes (BC 11.0% vs GC 4.0%, p<0.001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Distribution of types of CAM among users.

https://doi.org/10.1371/journal.pone.0285718.g003

Association between CAM use and QoL in BC and GC patients

Table 2 displays the mean value for each symptom scale, single item, and the global health status / QoL of the EORTC QLQ-C30 questionnaire. The proportion of high symptom severity is significantly higher among CAM users than among non-CAM users in the dimensions fatigue (p = 0.003), appetite loss (p = 0.032), and diarrhea (p = 0.007) in GC patients, in pain in both groups (BC p = 0.018, GC p = 0.032), and insomnia in BC patients only (p = 0.007). CAM users have (not significantly) better QoL scores in comparison to non-CAM users for nausea/vomiting, appetite loss (BC) and constipation (GC). The global health status / QoL among GC patients is significantly higher in non-CAM users than in CAM users (54.5 vs 47.9, p = 0.01).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. QoL (EORTC QLQ-C30) scores between CAM users and non-CAM users.

https://doi.org/10.1371/journal.pone.0285718.t002

The results of the logistic regression model are shown in Table 1, column B.

The likelihood of using CAM is increased in patients with higher educational status (OR = 1.57), middle income in comparison to low income (OR = 1.41), patients undergoing chemotherapy (OR = 1.66), being treated in a rehabilitation unit (OR = 1.58), and with higher distress levels (OR = 1.87). CAM is most likely to be used if the time since diagnosis is more than 12 months (OR = 2.01).

Discussion

In this analysis, we surveyed German BC and GC patients concerning their usage of CAM and related associations with QoL.

The prevalence of CAM use in the total study cohort was 46.5%. This is in line with other surveys of BC and GC patients [20, 28, 29]. However, there are several studies with higher prevalence rates, the reasons for which are manifold. In a review report [12] with GC patients from the USA, Europe, Israel, China, and Australia, the prevalence rates varied from 40.3% to 94.7%. According to the authors, the reasons for that broad range are: discrepancies in CAM definitions, diverse CAM modalities, varying research methods (self-report, interviews), as well as geographic and cultural differences in combination with varying attitudes of acceptance of CAM [12].

However, recent studies in particular show an increasing acceptance of CAM methods not only among cancer patients but also among practitioners. CAM is being integrated into conventional therapy, and specialized treatment units are being developed in hospitals in many places, a progression that is leading to increases in the prevalence of CAM use [17, 30, 31]. The difference in prevalence between BC and GC patients in our cohort is statistically significant (BC 48.6% vs GC 40.7%, p<0.015), similar to those observed by Fasching et al. [29] (BC 50.1% vs GC 44.0%, p = 0.058). The authors understood this difference to be due to BC patients belonging to a better organized network and thus having access to more information than GC patients [29]. Paepke et al. also showed a higher CAM use in BC patients (56.7%) compared to GC patients (43.3%), but did not suggest an explanation for why this is the case [32]. They also found that there was a high interest in CAM in the non-user group.

In contrast to the well-researched CAM prevalence, there is significantly less evidence on patient satisfaction regarding their CAM use. Our study shows a 60% satisfaction rate with CAM use ("very satisfied" and "satisfied"). These results have also been found at the same level in comparable studies [8, 20].

The most popular CAMs in both groups in our study were (1) vitamins/ minerals/ micronutrients, (2) homeopathy, and (3) mistletoe. These rankings are mostly in line with previous studies [7, 9, 30, 31], though our results reflected lower percentages for the use itself [17]. Neural therapy, kinesiology, and Bach-flowers therapy were hardly used. The CAM types/modalities in the studies from different countries and cultures are as varied as the respective prevalence of their use [6], a fact which limits their comparability.

CAM use is dependent on various sociodemographic and medical characteristics. In our study, the following characteristics were statistically significantly associated with CAM use: younger age (< 50 years), higher vocational education, higher monthly income, longer time since diagnosis, more advanced disease stage, treatment with radiotherapy, chemotherapy and hormone therapy, clinical setting (rehabilitation), and higher levels of distress. These associations are also reported in a number of other studies [6–10, 33]. In addition, there are studies that have failed to demonstrate any associations between sociodemographic or medical characteristics of CAM users [13, 14, 34]. Reasons for this are: frequently small sample sizes, differences in the availability of CAM, costs in different cultural settings and national health care systems.

Consistent with previous studies, the following predictors of CAM use we found were: higher education [7, 10, 11, 35], middle or high monthly income [10], time since diagnosis > 12 months [7, 8, 11].

However, other medical characteristics such as past/current treatment with chemotherapy, in-patient status in a rehabilitation setting, and higher distress levels are also associated with a significantly higher probability of CAM use in our study. Age also differs significantly between CAM users and non-CAM users in the bivariate analysis (p = 0.001). This statistical effect, however, is no longer detectable in the regression analysis, once it is influenced by the parameters of vocational education and distress level. Numerous other studies have demonstrated age as a predictor of CAM [7, 18, 28, 35]. Furthermore, in contrast to our results, other studies have also shown employment status [8] and stage of the disease [10, 18] to be predictors of CAM use. Often, however, no predictors of CAM use could be found [31, 36]. The CAM users in our study showed significantly lower QoL scores in many areas, i.e. a higher symptom burden compared to the non-users. In particular, these include pain in BC and GC patients, insomnia in BC patients, and fatigue, appetite loss, and diarrhea in GC patients. From a clinical point of view, this could mean that the motive for CAM use is often hope for symptom relief and to have an active way of coping with the disease. This is also supported by the results of Hwang et al., who showed more frequent CAM use in BC patients with greater symptom burden [11]. The global health status / QoL of GC patients was significantly higher in non-CAM users (54.5 vs 47.9, p = 0.01) in our study.

Due to the cross-sectional study design, no causal relationships can be derived here. Ben-Arye et al. examined CAM use at two different time points (baseline and a 6 to 12-week follow-up) and showed that most outcomes improved in the CAM intervention group [37]. Klafke et al. [38] were able to demonstrate improved QoL scores in fatigue, emotional functioning, and dyspnea among BC and GC CAM users at the end of chemotherapy compared to the start of chemotherapy. For BC patients in our study, no difference in global QoL was found between CAM users and non-CAM users, which is consistent with the results of most previous studies [10, 14, 16, 18, 21].

What could these findings mean for clinical practice?

Patients with breast and gynecological cancers have a broadly common clinical context with regard to diagnostics, therapy and after-care and are mostly treated by the same specialist. Therefore, both cancer types are discussed within the same framework.

The focus on a large patient group within our study demonstrates to which extent CAM methods are used and which differences can be found between the two types of cancer patients. For doctors as well as for clinical staff, it is very important to know which preferences patients have regarding their aftercare within the framework of oncological therapy. It helps all physicians to define and recommend the right additional medical treatment for each patient individually, which can be easily integrated into clinical care plans.

Although the number of studies on CAM use has multiplied over the last 10 years, evidence for the effects of CAM is still very sparse. Above all, longitudinal studies are necessary in order to evaluate causalities and also to derive better dose-response relationships. Adverse effects as well QoL scores respectively should be investigated regularly before, during and after the cancer therapy, preferably by using standardized questionnaires such as the EORTC QLQ-C30 or the QLQ-BR23. This approach would enable the evaluation of and recommendations for the use of CAM methods at a very individualized level. In particular, for the most popular CAM group of vitamins/minerals/micronutrients there are no dose specifications and the application should ideally be monitored by taking blood samples. Our and many other studies have shown that older patients in particular are less likely to use CAM [9, 12], yet high BMI and concomitant medication are also associated with a lower interest in integrative medicine. [28]. It represents a vicious circle as old age and high BMI increase the risk of chronic diseases and cancer itself. Furthermore, there a no effective screening strategies for gynecological malignancies such as Ovarian or endometrial cancer as shown in a recent study by Giannini et al [39]. CAM is to be considered here in terms of tertiary prevention in order to achieve the best possible QoL. The population will continue to age in the future, so that an increasing number of older people with comorbidities will be diagnosed with cancer compared to today. CAM use should also enable a strategy to improve the oncological outcome comparable to the preoperative evaluation of the frailty of gynecological cancer patients proposed by D`Oria et al [40].

Limitations and strengths of the study

There were some limitations of our study. First, it was a cross-sectional design and, therefore, no conclusion regarding causality with CAM use can be drawn. Second, we did not ask about participants’ motivation for CAM use and non-CAM use, and we did not ask about any previous patterns of CAM use prior to the respondent’s cancer diagnosis.

The strengths of this study are its large sample size and the multicenter study design.

By including patients in different treatment settings (acute care clinic, outpatient departments, rehabilitation centers), it is possible to examine CAM use throughout the treatment course. Because the methodology is identical, is it possible to directly compare both cancer types.

Conclusion

Though this is not the first study to examine patients with gynecological malignancies and their use of CAM (prevalences, modalities, characteristics, predictors, and associations with QoL), it does underline previous study results, especially in Germany. Creating uniform definitions for CAM and CAM modalities with regard to the comparability of study results is of enormous importance for further research in this area. Likewise, larger longitudinal studies should be conducted in the future to capture causality between CAM use and QoL.

Acknowledgments

We thank all healthcare teams who were involved in data collection across all of the local study centers.