Discussion
In this study we describe the state of the art of the existing scientific experimental evidence on the factors and determinants that influence patients’ adherence to treatment. As can be seen in the results section and the Supplementary Material, many studies have examined the effects of several factors and determinants on adherence to treatment. In particular regarding socioeconomic factors, most studies have considered determinants from the patients’ background or environment including their financial status, education, employment status, and living condition. Besides that significant associations have been found, these factors and determinants are difficult to modify or influence. Other factors studied relate to the existence of social support networks, which have been found to significantly affect adherence to treatment by several studies (Siregar and Andayani, 2020; Zullig et al., 2015; Reddy et al., 2017). Easily modifiable patient lifestyle factors have also been identified to have a significant contribution to adherence levels (Meggetto et al., 2017; Sieben et al., 2021; Llorca et al., 2021; Nascimento et al., 2016; Shankari et al., 2020). Multiple studies have explored the effect of those factors on adherence related to the patient-HCP relationship and to the different figures and institutions involved in healthcare. From these, the most relevant factor associated with adherence to treatment is the provision of education and follow-up to patients. In fact, several studies have identified its effect on the level of adherence to treatment (Tola et al., 2016; Bonetti et al., 2018; Hohmann et al., 2014; Hovland et al., 2020; Kamal et al., 2015; Ababneh et al., 2019; Wan et al., 2016; Asgari et al., 2021; Alfian et al., 2020; Jahn et al., 2014; Wang et al., 2020), although some studies did not find any significant relation. This evidence is highly relevant as it can guide future interventions and guidelines that can help improve patients’ adherence to treatment.
The characteristics of the treatment-related factors, the duration and its symptomatology have also been identified as having a strong influence on the patients’ adherence levels. Furthermore, not only the condition but also the treatment characteristics have been identified as strong influencers, with complexity and duration of the treatment being the major factors (Flicoteaux et al., 2017; Gillespie et al., 2014; Kuypers et al., 2013; Suffoletto et al., 2012; Wooldrich et al., 2015; Sieben et al., 2021; Kamal et al., 2015). The relevance of identifying treatment complexity as an adherence determinant may help HCPs when deciding on the treatment options available for a specific patient and the associated risk of non-adherence to such treatment.
Finally, there are some factors that are related to the physical and behavioral characteristics of the patients and their environment. Factors like age, gender and ethnicity that are unalterable for the treatment purpose have been identified by many studies as being associated with treatment adherence. However, not all studies agree on the direction of the effect of these factors, which indicates that the effects of these factors can be highly dependent on the study setting (e.g., type of disease, type of treatment, intervention, participants included). Other factors identified as modifiers of the adherence levels were factors related to the patients’ health literacy and pre-existing beliefs and concerns about the condition and the treatment (outcomes) (Crowley et al., 2014; Nieuwkerk et al., 2012; Crowley et al., 2012; Beckers et al., 2012; Al-Haj Mohd et al., 2016; Shankari et al., 2020; Siregar and Andayani, 2020; Llorca et al., 2021). Other relevant factors are those related to the patients’ lifestyle, their self-efficacy and planning abilities (Gillespie et al., 2014; Mugo et al., 2014; Wan, 2016; Llorca et al., 2021; Shankari et al., 2020; Kuypers et al., 2013; Crowley et al. 2014). The identification of these factors related to the competences of the patients, their behaviors, and psychosocial factors is highly relevant to better understand a patient’s behavior towards recommended treatments and to better design approaches to improve the patient’s adherence levels.
It should be noted, however, that this study has some limitations. First, the eligibility criteria limited the search to those studies published in the last decade in English. Still, most studies nowadays are published in English, and we see the studies do not show a bias towards studies based on English-speaking regions. Second, our SR has shown that regarding adherence to treatment, most studies focus on adherence to medication and do not include additional treatment options, such as lifestyle changes, which are necessary in most cases. Subsequently, we also see that most of the studies rely only on self-reported data (N=31), a small number of studies used pill counts (N=7) or devices on medication (N=7), and only 2 used biochemical analytic data. Furthermore, most studies have used only one type of adherence measurement, making it difficult to compare the outcomes. Especially considering that differences in measuring methodology may lead to differences in the assessment of adherence levels. Importantly, the fact that self-reported data carries the biases of recall and social desirability, along with its lack of granularity and general overestimation of adherence, is a limitation for the accuracy and precision of the data collected. Third, none of the studies have included patient adherence to treatment across the most common diseases (e.g., cardiovascular, oncology, immunology, neurology, endocrinology, and rare disease), making a comprehensive understanding of patient adherence difficult. In fact, only a limited number of the included studies covered multiple of these condition areas, and most focused only on one area. Another important lesson learned is that most of the studies consider participants from one country only, which makes it challenging to assess generalizability of the obtained results to other countries or regions where socioeconomic and healthcare system-related factors might significantly differ. Remarkably, none of the studies included the cost of treatment in their analyses, although this is an important determinant of adherence to treatment, considering the importance of the socioeconomic factors in selected studies. Fourth, regarding the review process, having such a broad topic and scope (including several kinds of conditions, treatments, measures of adherence, etc.) challenges the proper feeding of the ASReview tool. This limitation has been overcome by performing several additional training rounds before getting the final prioritization algorithm. Lastly, most of the literature studying factors influencing adherence to treatment relies on patient self-reported data, which, as discussed above, carries its own biases. These are vital lessons learned for future steps in scientific research in patient adherence to treatment.