Scientific Papers

Impact of interventions on the quality of life of cancer patients: a systematic review and meta-analysis of longitudinal research | Health and Quality of Life Outcomes

The quality of full-text paper

In this study, two types of paper quality assessment scales were applied, and we found that 122 included papers have good and fair quality in general. As regards The Newcastle-Ottawa Scale is useful to evaluate longitudinal studies with non-randomized control trial design. With this scale, we found that nearly 40% of included studies were good while 60% were fair. The primary reasons are the lack of information on the selection, exposure, and outcomes sections. Some studies did not have specific criteria for selecting subjects, or reasons for selecting subjects in the selection. In addition, studies did not provide detailed information on how outcomes and exposure were collected and evaluated at the end of the process. The results of our study are similar to the systematic review of chemotherapy and surgery’s impact on the quality of life of breast cancer patients in 2022. Among 26 studies collected, 34% were good and 66% were fair [27]. However, another systematic review and meta-analysis in 2020 researched on survival rate in colorectal cancer and they found that 60% of the collected paper was at a good rate and 40% at a fair rate [28]. Although the majority of included studies getting medium or high quality, there are still some studies that need to be considered more carefully when developing a research design, analysing the data, and presenting the results.

The other type of quality assessment scale, the Jadad Scale, determined that approximately 70% of randomized control trial studies had high quality and only one-third of them had low quality. When we scored the included studies, we found that some of the articles lack information in the description of randomized and blinding criteria. They did not specify the reasons for selection or the method of selecting research subjects into random groups or did not specify the method for blinding criteria. Some articles did not mention whether they blinded the subject. The results of our study are higher than a systematic review in 2013 that researched the effectiveness of palliative care in cancer patients. The reason for this difference may be that the above author’s study only focused on evaluating randomized control trials (RCTs) conducted in China without other countries [29]. Another systematic review conducted in 2017 also used Jada’s scale to evaluate the quality of RCTs and found that there were 10/18 (55.6%) studies with Good quality which is lower than our study [30]. However, the authors only focused on evaluating studies that applied an intervention or pain management and were conducted in Africa, Europe, and North America.

Although the systematic review and meta-analysis mentioned above are conducted in different sets of studies, the quality evaluation of included studies may be essential. It might give information about the current trends in research design and present, especially with the rapidly increasing amount of research being published. We believed that it is essential to apply scales such as the NewCastle-Otawa and Jadad Scale to evaluate them carefully.

The characteristic of included longitudinal research

Of the 122 studies selected for meta-regression analysis, we found that nearly 50% of the studies were designed with a 6–12 month follow-up period. At this point, if the follow-up time is too short, it may not fully reflect the change in the patient’s QoL indicators. However, if the follow-up time is too long, it can lead to the loss of patients because of cancer and especially metastatic cancer, the patient’s survival rate will change from time to time. The loss of patients can lead to biases in the evaluation of longitudinal research. However, in a long time of follow-up, they may have a better assessment of the change in the health utility of cancer patients because they will not miss the decrease in the quality of life of cancer patients in their last days. Consequently, these follow-up assessments are often difficult to achieve high accuracy avoiding biases.

In this research, we counted about three hundred single tools used to evaluate different aspects of the health utility of cancer patients. These tools can be classified into eight main groups named: EQ-5D, VAS, EORTC-QLQ, FACT, FLIC, SF36, QLI, and others. The most used scale was EORTC-QLQ. However, the EORTC scale can be designed with many other subgroups such as EORTC QLQ C36, EORTC QLQ HN37, EORTC QLQ HN35, EORTC QLQ OV28, EORTC QLQ BR23, EORTC QLQ BR38, and EORTC QLQ LC13 which are used to assess different types of cancer. We found that another group of authors who conducted a systematic review of 13 articles on the health-related QoL of prostate cancer patients claimed that 7/13 articles used the EORTC scales [31]. In addition, another research in 2021 also assessed the impact of physical therapy on oesophageal cancer patients with all the included studies using the EORTC scale [32]. This result is different from our study but is due to the huge difference in the number of studies evaluated. Although the EORTC-QLQ scale is very useful, it is necessary to compare different scales, especially with some other important scales such as EQ-5D, VAS, FACT, and SF- 36.

On the other hand, FACT is another important scale with nearly six hundred measurements made with FACT in the 122 included studies. This scale also has several subgroups and tools, which can be modified for different types of cancer intervention impact assessment. In another study, besides the EORTC scale, FACT is also commonly used with a ratio of 26/43 of the researches [16]. Therefore, FACT is one of the most common scales which may be needed more in-depth research.

Regarding the contribution of different countries in this research field, Sweden ranked at top of the list with nearly a thousand of these tools. Moreover, the data showed that America and China still had a high number of QoL measurements, followed by other Western European countries and Canada. It is difficult to compare our results with previous publications because our research was not limited to any region or country in the world.

In addition, there was a significant difference between the pharmacological and non-pharmaceutical interventions. The non-pharmaceutical intervention has only been the focus of research in recent years. However, this intervention group still played an important role in the improvement of the well-being of cancer patients. As for the type of cancer, lung cancer was the highest rank with more than seven hundred health utility measurements. It was followed by breast cancer, head, and neck cancer, prostate cancer, and other types. These were all cancers with high incidence rates with several burdens on the QoL of cancer patients. Therefore, the majority of research conducted so far often focused on these diseases. In a systematic review of the impact of exercise on patients with various cancers conducted in 2017, the most frequent cancers, in descending order, were found to be: breast cancer, gastrointestinal, head and neck cancer, cancer, endometrial and ovarian cancer, prostate cancer, lung cancer, blood cancer, and others. There is a remarkable difference in the rank of cancer types compared to our study, but this difference is due to our comprehensive approach to types of cancer intervention [33]. In this systematic review, we realized that some cancers such as breast cancer, lung cancer, and prostate cancer had a higher research focus. Whereas some other types of cancer are uncommon, the sample size is small, and the research and evaluation can appear less often. This is similar to the group of interventions, the studies mainly only evaluated some common methods such as screening, chemotherapy, radiotherapy, surgery, palliative care, and a combination of them. Thus, more future studies are needed to evaluate more about some rare cancer types and new modern interventions.

The adjusted effect models

Our group is one of the first to apply random effect models to investigate the change in the health utility point of cancer patients. We evaluated and compared the impact of the group of quality of life measures, and the group of measures against other factors such as types of cancer, types of intervention, country types, and study design. The results showed that there are statistically significant changes when comparing groups of health utility measurement tools, cancer types, and interventions. Among them, we found some prominent scales such as VAS. This scale exhibited some differences when included in the comparison models. In a systematic review of self-report instruments for the measurement of anxiety in hospitalized children with cancer in 2021, the authors also recommend that the VAS scale be combined with other self-report scales [34]. Besides, commonly used scales including EORTC and FACT have no statistically significant difference when included in the models compared to the scale EQ-5D. However, the role of these scales in cancer intervention research is essential. In the systematic review in 2011, the authors suggested that the EORTC and FACT scale rank at the top of the list for impact scales [35]. This research suggested that further evaluation of the VAS scale will need to be done to determine its effectiveness in scoring patients’ QoL.

Besides, several cancer groups including acute myeloid leukemia, brain cancer, colorectal cancer, oesophageal or gastroesophageal junction cancer, Hodgkin lymphoma, and prostate cancer had a positive statistically significant change. On the other hand, cervical cancer and kidney cancer had negative statistically significant changes in model of type of cancer. Cancer patients in these groups are often monitored for changes in the QoL point continuously over a certain period along with the impact of interventions. From there, scientists can further evaluate changes in QoL points with different influencing factors.

In addition, when evaluating the interventions, we found that the radiotherapy group, screening, combined supportive care and chemotherapy, palliative care group, combined curative and palliative care group and physical exercise group elicited significant changes in the quality of life of patients. This result emphasizes the need for close coordination between pharmacological and non-pharmaceutical interventions to improve the quality of life for cancer patients. Curative care interventions still play an important role to help the patient’s QoL, especially radiotherapy. However, the results also indicated that major changes were often concentrated in groups that combine multiple approaches between pharmaceutical and non-pharmacological interventions. Indeed, best supportive care combined with chemotherapy and curative combined with palliative care groups showed statistically significant changes in the QoL score when apply to cancer patients during the time of follow-up.

A systematic review in 2017 indicated that patients with comorbidities and chemotherapy had decreased HRQOL after treatment [16], however, the authors suggested that the combination of additional palliative care and social support can improve the quality of life of cancer patients. While we found significant differences in QoL scores when comparing cancer types, interventions, and influencing factors such as country types and study design when applying the different scales. Nonetheless, there is no significant statistic when analysing by country type and study designs. Thus, these factors may not be the primary factors affecting changes in health utility scores.

However, there were some limitations in our research, including the stage of cancer, the demographic of participants, and several types of high-technology intervention that were not analysed in our results such as targeted drugs, new gene sequencing, and immunotherapy. Thus, future researchers should consider these aspects and conduct more research with different types of biostatistical analysis models. The above factors can play an important role in the change of the health utility point of cancer patients under different interventions. However, the random effect models were chosen because we believe that it may be effective on analysing the huge amount of data and the variety of vital affecting factors. In addition, the sensitivity was not analysed as one of the limitation of us. However, it should be considered carefully in the further research.

Source link