During the first wave of the COVID-19 pandemic, a global lockdown was imposed in response to the rapid spread of the virus; the government’s regulations or restriction measures to curtail virus spread may have an enormous impact on people’s daily lives (1, 2). It is self-evident that the implementation of quarantine measures, lockdown, and social distancing protocols also led to some negative repercussions, which include a significant meltdown in economic growth, an increase in unemployment or financial insecurity, a rise in the cost of living, and a severe impact on health service utilization among the population (3–5).

With the continued containment and control of the COVID-19 pandemic, the general population also experienced unintended potential health risks arising from the restriction of routine healthcare services, specifically, the consequences of the pandemic for the health of these non-COVID-19 patients, including delayed timely detection and treatment, avoided and delayed emergency department visits, unmet healthcare needs and increased rates of exacerbation as resources were reallocated to urgent care for COVID-19 patients (6), researches noted the COVID-19 pandemic has adversely affected the healthcare utilization of the population worldwide, and situation of restricted health service utilization is alarming (7, 8). A national longitudinal study from China found the most considerable negative impact of the COVID-19 pandemic on health services utilization was observed between Jan 2020 and Apr 2020, with approximately a 32% reduction in hospitals, a 22% reduction in community health centers, and 27% reduction in township centers (9). Studies from Europe have claimed that during the COVID-19 pandemic, individuals with acute myocardial infarction, stroke, heart failure, and other chronic cardiovascular diseases experienced a significant reduction (40%) in hospitalization rates and emergency department visits compared to baseline (10). Globally, a large study indicated that global healthcare utilization showed a median decline of 37.2% between the pandemic and pre-pandemic periods, ranging from 19.8 to 50.5% (8). In short, the fear of infection and reduced availability of medical services have driven down non-COVID-19 healthcare utilization. The demand of the general population, particularly vulnerable groups like chronic patients, to improve their health status is a difficult task under normal circumstances, the obvious barriers imposed by the COVID-19 pandemic have considerably increased this difficulty. Meanwhile, considering the healthcare system supply is limited, it appears to be a moral dilemma between allocating medical resources to contain the spread of the virus or providing adequate healthcare service during the pandemic. Nevertheless, it also indirectly drives us to be more concerned about exploring strategies and measures to address the restricted health service utilization more effectively.

Since the outbreak of the COVID-19 pandemic, the infection spread has come in waves in China, restrictions on health services were occasional (9), and growing research has focused on self-care behavior and recommended it as a response strategy to tackle the potential health harms arising from restricted health service utilization (11, 12). Drawing on the various relevant literature, the restricted health service utilization (RHSU) not only hinders surveillance of health status but also will likely lead to some worse outcomes if it continues, such as unhealthy lifestyles, treatment interruption, and worsening of chronic disease symptoms (13), all of which could lead to serious health problems. To alleviate this potential health crisis, the people with intentions to maintain health and those with chronic diseases preferred to adopt non-therapeutic measures, including physical activities, dietary habits, self-care monitoring and online medical consultation after the outbreak (14–16). Previous studies found that several types of health service utilization restrictions were associated with self-care activities, such as both primary and specialized care (17), personal care aides (18), primary mental health care service (19) and chronic disease treatment (20, 21). Drawing upon evidence from related research, the RHSU might directly or indirectly motivate the general population to compromise in daily choices like consumption habits, diet, exercise, self-care activity and medical consultation, and most people had mitigated potential health harms through these behavioral changes (14–16, 22). For instance, a study of stroke survivors stated that self-care behavior such as physical activity, diet, weight control, smoking cessation, and abstinence from alcohol could help those with restricted health services to reduce systolic and diastolic blood pressure and decrease the occurrence of complications such as hyperglycemia and diabetes and keep disease stable (23). Medical practitioners have also suggested that adopting self-care activities and improved self-care behaviors may play a critical role in maintaining health or preventing immediate and subsequent complications during the COVID-19 pandemic (24).

Specifically, self-care behavior is recognized as an essential and valuable behavior because it emphasizes the positive role of people in maintaining their health, and common individual self-care activities include engaging to improve health status, prevent disease, limiting illness, and regain health (25), the beneficial effects of self-care include improved well-being and lower morbidity, mortality, and healthcare costs (21). Previous research has revealed the complexity of self-care and illustrated the wide variety of factors that influence the decisions individuals make about engaging in self-care (13–16); it also investigated the difficulty performing self-care among special populations (e.g., multiple chronic conditions, severe mental illness, low health literacy) (26); and sustained evolving technology to enable individuals to manage their conditions and improve the efficiency of self-care (27), the multiple disciplines are actively studying self-care and contributing variable knowledge to the topic nowadays (28). However, the current literature suggests that self-care is still under-appreciated and insufficiently understood; there are many challenges to the prevalence of self-care behavior (26). Thus, there is a need to explore the mechanisms underlying self-care behavior further to drive relevant policy development, especially for those countries and regions with a growing burden on health service systems.

Since March 2020, the World Health Organization declared COVID-19 a global pandemic, and governments around the world imposed restrictions on the use of hospitals and outpatient services, eliminated all elective, routine and non-emergency patient procedures, implemented stricter physical distancing measures and transitioned to remote care to reallocate resources to the urgent care of patients with COVID-19 (42), which substantially disrupted individuals’ routine healthcare utilization. Globally, the Chinese government’s endeavor to contain the spread of the COVID-19 pandemic has been widely praised, which may have contributed to a more severe influence on individuals’ routine healthcare utilization during the early stages of the outbreak compared to other countries (9). To the best of our knowledge, this is the first study to explore the impact of health service utilization on self-care behavior among Chinese adult populations during the early COVID-19 pandemic based on a large sample covering 31 provinces in China. Furthermore, our study used two dimensions with four indicators to measure self-care behavior, explore the relationship between RUSH and self-care behavior, and examine vulnerable populations and subgroup differences through multilevel regression and subgroup analyses. Our findings from the COVID-19 pandemic may further elucidate the future policy development of self-care behavior in China, provide appropriate interventions to mitigate the health risks caused by RHSU and contribute to alleviating the burden on the public health system and its long-term benefits.

Firstly, the multivariate results supported much of what we anticipated earlier: the presence of RHSU was positively associated with all four kinds of self-care behavior in this study, and these associations remained significant even after adjusting for individual, environmental, and risk-perceptive control variables. Similar to the results of studies before (like medical services supply decline) and after the outbreak of the COVID-19 pandemic, these researchers also observed the association between social inequalities, pandemic-related changes, and individual health behavior (41–43). For instance, when individuals have potential health risks (including reduced accessibility to healthcare or deterioration of health status), this may stimulate the self-care activities and behaviors that were aimed at preventing or reducing health risks and optimizing health and quality of life (22, 44). Also, evidence from Anderson’s health behavior model likewise supports a relationship between the use of preventive health services and self-care ability in daily life (46). Regarding autonomous behavior, previous studies claim that there is an association between decreased healthcare utilization and increased leisure-time physical activity; this association remains after adjustment for socio-economic confounders (17), which is consistent with our findings. Research with adults during the COVID-19 pandemic suggests that self-isolation at home due to lockdown is associated with a lower level of physical activity and modifications in eating behavior (47, 48). Also, a classic literature review elucidated that individual age, health literacy, and self-rated health could indirectly influence the association between healthcare utilization and physical activity through mediation analyses (49). Based on the discussion above, we predict that the relationship between self-care behavior and RHSU will persist even after the COVID-19 pandemic subsides.

Secondly, even if there is a willingness to practice positive self-care behavior, some vulnerable people still perform worse in familiarizing self-care activities and building self-care behavior due to physical and psychological factors (26), which if ignored could make it difficult to implement the entire strategy of reducing the burden on public health through self-care behavior. For example, previous studies indicated that 30–49-year-old adults, those with a higher level of education, and those who were employed and had a high income were more knowledgeable and better able to take appropriate measures to prevent the spread of COVID-19 (35, 50). However, our study observed that participants in the survey had generally high scores on prevention behaviors related to COVID-19, and these differences between populations were not apparent compared to other studies, which may be attributed to the Chinese government’s extensive publicity and appropriate supervision of interrupting the spread of the disease in the early stages of the COVID-19 pandemic (51).

Similarly, our study indicated significant correlations in online medical consultation across age, gender, income, married status, health status, and infection risk perception, those findings align with previous studies (27, 52). It is not surprising that social distance, isolation, and hospital restrictions forced chronic patients to re-organize their routine care through temporary in-person visits during the pandemic, this led to a widespread and significant increase in telemedicine utilization (3). Also, patients with chronic disease may avoid in-person visits to hospitals, clinics, and emergency departments for fear of exposure to potential COVID-19-infected patients (3, 53), thereby preferring to implement telemedicine or seek online medical advice. However, our results showed that middle-aged and older adults showed relatively weak performance in medical consultation compared to younger adults. In fact, despite research examining age differences in self-care during the COVID-19 pandemic has been inconclusive, previous studies on the digital divide demonstrated that people aged above 65 years are less likely than the younger generation to have had the chance to familiarize themselves with ICT either at school or at work, combined with the cognitive, motor, digital gap and sensory decline associated with aging, and older adults face more barriers to and challenges in using online technology for health than their younger counterparts (54). The present studies claimed that this digital divide created by digital technologies might widen social inequalities by alienating disadvantaged groups that do not have access to digital resources (55, 56). Thus, we recommend that particular attention be paid to the older adult when future discussions on critical issues surrounding the promotion of online medical consultations include quality of care (57), communication and language barriers (58), and patient satisfaction (43)^.

Thirdly, our study also extended the existing literature by the correlations between restricted health service utilization and self-care behavior, which were quantified and compared by subgroup analysis. Based on this, we found that external and intrinsic risk factors had a significant effect on some type of self-care behavior. For example, residents living in the higher infection-risk areas were more likely to report above-average online medical consultation scores or engage in physical activities when they experienced RHSU. The strength of this association varies significantly with the infection risk level of the living area. These findings are also consistent with the argument for behavioral change theory since high infection-risk living environments created additional barriers (e.g., extra healthcare costs, and psychological stress) to illness prevention and health maintenance. Therefore, except for inequities in health service utilization, the risk perception (41) and actual infection risk related to the pandemic (44) may be critical indicators of health behavior choices. A possible explanation from psychology is that the higher the risk an individual perceives, the more motivated they might be to engage in protective behaviors (53). Despite this, both external and intrinsic risk factors need to be considered for personal characteristics, previous research highlighted that not everyone responds to health risks similarly and that risk perception alone does not explain health behavior (59). For instance, the self-care behavior of chronic disease patients is associated with a high perceived susceptibility to disorders; they may tend to avoid in-person visits to hospitals, clinics, and emergency departments for fear of exposure to potential COVID-19 (43). Likewise, older adults are more vulnerable to COVID-19 infection, have a worse prognosis after infection, and have a higher risk of getting one or more non-communicable diseases, so that they may experience heightened levels of instilled fear of COVID-19 exposure during in-person medical services (3). In general, caution must be exercised in the following discussion of these findings.

In contrast, no significant associations were found in the model of weight control among subgroups analysis, which means that the strength of the association between RHSU and weight control was not affected by age, chronic diseases, and actual infection risk level of the respondent’s residence. Still, we found several reasons that may confound our findings by searching the literature. A study from Obesity has suggested that social closure measures may have a wide-ranging effect, making it more difficult for many people to adopt weight gain protective behaviors (60). It was widespread for people to experience barriers to diet and healthy eating during the COVID-19 pandemic lockdown (43, 47) (e.g., food shortage, lacking motivation and control around food). Also, some mental health issues, such as anxiety, stress, and poor mood, were found to be risk factors for obesity during the pandemic (61). Collectively, the factors mentioned above may make it more difficult to distinguish differences in weight control behaviors among populations.

At last, given that individuals who experienced restricted health service utilization were more actively engaged in self-care behavior, our study examines the interaction of health service inequalities and potential risk factors on self-care behavior during the pandemic. It highlights the necessity of promoting self-care behavior when in response to health service restrictions, especially for vulnerable individuals (62). Meanwhile, in view of the various benefits of self-care behavior, we need more comprehensive policies to encourage individuals to develop self-care behavior. Both primary healthcare institutions and community service organizations should strengthen the strategic support related to self-care promotion and implementation for the public regarding personal health literacy development, public healthcare services, online healthcare services, continuing medical education, etc.

From the perspective of health care system reform, addressing current barriers around self-care in terms of applicability, developmental disequilibrium, standardization, supervision, service coverage, and digital divide caused by technological advances will help more people maintain their health through self-care activities effectively; it not only effectively responds the issue of restricted health service utilization, but also contributes to alleviate the burden on the public health system and its long-term benefit (63). Indubitably, there is a pressing need to strengthen the development of self-care systems in China.

Our study includes several fundamental limitations that must be acknowledged and addressed in future studies. Given the cross-sectional nature of this study, one of the main limitations was that there was no baseline response rate before the pandemic and no available data on participants’ previous self-care behavior; neither can make definitive statements about causality in regression analyses. Second, this study assessed outcome variables by employing a few single-item scales; also, only three dimensions were collected to describe the performance of self-care behavior. Therefore, future research should determine variables more comprehensively by using multiple scales to provide more conclusive evidence on the predictive validity of self-care behavior. Third, participants were recruited using a snowball sampling method through social media; the advantage of this method is that a large number of samples can be collected quickly. However, many participants were well-educated and below 40 years old, leading to a particular bias in the results, which made it difficult to identify more subgroups with significant differences. Fourth, the subgroup analyses in this study did not use more demographic categories, such as region, economic status, and education level; these subgroup differences require deeper exploration in the future.

However, this study has some innovative findings. First, our findings extend the existing literature by exploring the impact of restricted health service utilization on self-care behavior in a large sample covering all provinces in mainland China. Based on the RAA theory, our study may find some motivations and factors for self-care behavior change; these may provide some references for promoting the effectiveness of self-care behavior to reduce the burden on the health service system. Second, in contrast to previous studies mainly focusing on the impact of inequalities in health service accessibility or quality on self-care behavior in the context of regional economic disparities, our study contributes to the existing knowledge base by investigating the relationship between restricted health service utilization and self-care behavior during a large scale infectious disease crisis. Third, we also take into account differences in risk perception when exploring the impact of inequalities in health service utilization on the outcome variables, so we tried to describe and compare the association between the RHSU and self-care behavior among the sex, age, chronic disease and high, middle or low risk of infections living area subgroups, it is effective in terms of filling the gap in the relevant literature.

The COVID-19 pandemic dramatically impacted health service utilization across China (9), and this study sheds light on the restricted health service utilization may predict more positive self-care behavior during the pandemic and the differential presentation of this association between subgroups of sex, age, chronic disease, and actual infection risk level of the residence area. Based on our results and current research findings, we believe that the correlation between restricted health service utilization and self-care behavior will persist as the COVID-19 pandemic subsides. Thus, we need further research into the mechanisms of self-care behavior, as well as continuing to address self-care knowledge gaps and improve outcomes. Given the many existing challenges to our vision, it is necessary to drive the development of policies related to self-care behavior to raise self-care as a vital element in general health and healthcare, which not only effectively responds the adverse health outcomes from restricted health service utilization and future public health crisis, but also contributes to alleviating the burden on the healthcare system and its long-term benefits.

The data analyzed in this study is subject to the following licenses/restrictions: the datasets presented in this article are not readily available because ethics restrictions. Requests to access these datasets should be directed to youfawang@gmail.com.

The studies involving humans were approved by the Institutional Review Committees of the Xi’an Jiaotong University, China (approval number 2020-1172). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin.

ZW: Conceptualization, Data curation, Formal analysis, Resources, Software, Writing – original draft, Writing – review & editing. ZZ: Data curation, Formal analysis, Funding acquisition, Writing – review & editing, Supervision. GL: Investigation, Methodology, Project administration, Writing – original draft. JL: Resources, Writing – review & editing, Formal analysis, Investigation. XZ: Methodology, Writing – review & editing, Resources, Validation. XF: Data curation, Investigation, Writing – review & editing, Methodology. SL: Conceptualization, Data curation, Investigation, Writing – review & editing. YW: Conceptualization, Data curation, Investigation, Writing – review & editing.