key: cord-0811027-06y7qbtj
authors: Xiao, Hong; Dai, Xiaochen; Wagenaar, Bradley H.; Liu, Fang; Augusto, Orvalho; Guo, Yan; Unger, Joseph M
title: The impact of the COVID-19 pandemic on health services utilization in China: Time-series analyses for 2016–2020
date: 2021-03-24
journal: Lancet Reg Health West Pac
DOI: 10.1016/j.lanwpc.2021.100122
sha: a94191068bf6c5ef5e04117eb8849d0bd152ac7e
doc_id: 811027
cord_uid: 06y7qbtj

BACKGROUND: The aim of this study is to quantify the effects of the SARS-CoV-2 pandemic on health services utilization in China using over four years of routine health information system data. METHODS: We conducted a retrospective observational cohort study of health services utilization from health facilities at all levels in all provinces of mainland China. We analyzed monthly all-cause health facility visits and inpatient volume in health facilities before and during the SARS-CoV-2 outbreak using nationwide routine health information system data from January 2016 to June 2020. We used interrupted time series analyses and segmented negative binomial regression to examine changes in healthcare utilization attributable to the pandemic. Stratified analyses by facility type and by provincial Human Development Index (HDI) – an area-level measure of socioeconomic status – were conducted to assess potential heterogeneity in effects. FINDINGS: In the months before the SARS-CoV-2 outbreak, a positive secular trend in patterns of healthcare utilization was observed. After the SARS-CoV-2 outbreak, we noted statistically significant decreases in all indicators, with all indicators achieving their nadir in February 2020. The magnitude of decline in February ranged from 63% (95% CI 61–65%; p<0•0001) in all-cause visits at hospitals in regions with high HDI and 71% (95% CI 70–72%; p<0•0001) in all-cause visits at primary care clinics to 33% (95% CI 24–42%; p<0•0001) in inpatient volume and 10% (95% CI 3–17%; p = 0•0076) in all-cause visits at township health centers (THC) in regions with low HDI. The reduction in health facility visits was greater than that in the number of outpatients discharged (51% versus 48%; p<0•0079). The reductions in both health facility visits and inpatient volume were greater in hospitals than in primary health care facilities (p<0•0001) and greater in developed regions than in underdeveloped regions (p<0•0001). Following the nadir of health services utilization in February 2020, all indicators showed statistically significant increases. However, even by June 2020, nearly all indicators except outpatient and inpatient volume in regions with low HDI and inpatient volume in private hospitals had not achieved their pre-SARS-COV-2 forecasted levels. In total, we estimated cumulative losses of 1020.5 (95% CI 951.2- 1089.4; P<0.0001) million or 23.9% (95% CI 22.5–25.2%; P<0.0001) health facility visits, and 28.9 (95% CI 26.1–31.6; P<0.0001) million or 21.6% (95% CI 19.7–23.4%; P<0.0001) inpatients as of June 2020. INTERPRETATION: Inpatient and outpatient health services utilization in China declined significantly after the SARS-CoV-2 outbreak, likely due to changes in patient and provider behaviors, suspension of health facilities or their non-emergency services, massive mobility restrictions, and the potential reduction in the risk of non-SARS-COV-2 diseases. All indicators rebounded beginning in March but most had not recovered to their pre-SARS-COV-2 levels as of June 2020. FUNDING: The National Natural Science Foundation of China (No. 72042014).

Responding to an outbreak of SARS-COV-2 in December 2019, China implemented a range of behavioral and clinical interventions to mitigate the epidemic in China and worldwide. These unprecedented interventions included large-scale diagnostic testing, rapid isolation of suspected and confirmed cases and their contacts, comprehensive contact tracing, and large-scale stringent restrictions on mobility.

In addition to banning travel to and from Wuhan city on January 23rd 2020, all provinces, municipalities and autonomous regions in mainland China subsequently activated the Level-I response to public health emergencies, the highest level of such responses, within one week [1] [2] [3] [4] [5] . As part of the emergency response, consistent guidance on managing the pandemic was implemented, intracity public transport suspended, intercity travel prohibited, public gathering banned, schools and entertainment venues closed, and information widely disseminated in a timely manner [2 , 3 , 6-8] .

Healthcare utilization has repeatedly been observed to decrease during pandemics, due to such factors as mobility restrictions, social distancing measures, and fears of contracting the virus within health facilities, as patients and healthcare providers defer or forego routine healthcare, especially elective and preventive visits [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] . The magnitude of the impact of the pandemic have varied markedly according to the type of healthcare service, location, and type of facility [11 , 12 , 14-16 , 21-25] , effects that have exacerbated existing inequities in the health system. In a recent example, analyses using routine health information system data from Guinea, Liberia and Sierra Leone showed substantial reductions in the delivery of maternal, child and reproductive health services, disruptions in HIV and tuberculosis testing, and large-scale collapse of vaccine and malaria case management programs during the 2014-2015 Ebola virus disease outbreak [9 , 17 , 26 , 27] . Most primary health care indicators in Liberia recovered to pre-Ebola levels one year after the end of the outbreak, although a similar rapid recovery was not observed in Guinea [9 , 17] . Significant reductions in admissions, diagnoses and treatments of stroke and acute myocardial infarction have been reported in China, France, Italy and America early in the COVID-19 pandemic [12 , 14 , 15 , 22 , 28] . Notable decreases in enrollment in non-SARS-COV-2 clinical trials were found despite proactive steps taken to allow for greater flexibility such as remote consent and virtual visits [29] . To our knowledge, no study has comprehensively examined the impact of the COVID-19 pandemic on national healthcare services utilization and the resumption of services after the pandemic in China.

The aim of the present study is to examine the impact of the SARS-COV-2 outbreak on outpatient and inpatient services utilization across China. We also investigated whether trends in utilization differed by area-level socioeconomic status (SES), category of health facility and treatment setting.

We extracted all available data for monthly facility-based health services utilization from the routine health information system of the Center for Health Statistics and Information, National Health Commission of China for the period from January 2016, to June 2020. Since 2007, all types of health facilities at all levels except for village clinics and primary care clinics have been required to directly report monthly health services utilization to the Health Commissions of the 31 provinces, municipality and autonomous regions in mainland China (except Hong Kong and Macau) via the provincial level electronic direct-reporting platforms. Countylevel Health Bureaus are responsible for collecting utilization data from village clinics and primary care clinics and then reporting the data directly to Provincial Health Commissions on a monthly basis. Provincial Health Commissions verify and review the data at the primary level, investigate missing report, aggregate the data and report to the National Health Commission. The National Health Commission oversees and checks the quality of data at the secondary level, and also provides technical training and support to staff at the provincial and county level involved in the reporting process [30] ). The primary indicators were the monthly number of health facility visits (outpatient and emergency department visits) and number of inpatients discharged by region and/or by health facility type nationwide. Monthly utilization counts for hospital or township health center (THC), a subtype of primary health care facility (PHCF), were available at the provincial level. Data for hospitals (public/private; first/secondary/tertiary/undefined level), other subtypes of PHCF (community health center, primary care clinic and village clinic) and other health facilities were aggregated at the national level. Subnational level human development index (HDI), a summary area-level measure of life expectancy, education and income level, was retrieved from the China National Human Development Report 2019 to reflect regional-level SES. A key advantage of examining an area-level measure in this context is its utility in providing evidence to help guide community-level interventions and policies. Each province, municipality and autonomous region was categorized into one of the three (high, middle and low) SES groups based on its HDI. Yearly urban and rural population estimates by region were extracted from China Statistical Yearbooks. This study was exempt from institutional review oversight since the data are publicly available and aggregated at population level.

Our aim was to examine a series of three interrelated research questions, including: (1) Was the SARS-CoV-2 outbreak associated with changes in outpatient and inpatient services utilization? We hypothesized that the outbreak resulted in meaningful reductions in utilization; (2) Did potential changes in healthcare services utilization due to the SARS-CoV-2 outbreak differ by regional-level SES, health facility type and treatment setting within health facility? We hypothesized that heterogeneity in the magnitude of relative changes in utilization existed across regions with different SES, various types of health facilities at different levels, and between outpatient and inpatient care setting; (3) Did changes in healthcare services utilization associated with the SARS-CoV-2 outbreak return to pre-pandemic patterns after the pandemic, and did patterns of recovery differ by regional SES, facility type and treatment setting? We hypothesized that healthcare utilization rates recovered after the peak of the pandemic in China-though to lower rates than before the pandemic -and that but the extent of recovery differed by regional SES, facility type and treatment setting.

We conducted an interrupted time series analysis to examine the impact of SARS-COV-2 on healthcare services utilization [31 , 32] . The impact was modeled using a segmented negative binomial regression with pre-SARS-COV-2 trends (January 2016 -December 2019) and indicator variables for each month in post-SARS-COV-2 period (January-June 2020). Due to known heterogeneity in the levels as well as the long-term trends of healthcare utilization across provinces, we employed a mixed-effects model with random intercepts and random slopes over time for province in subnational analyses. To adjust for observed seasonal patterns and the effect of the spring festival, we included fixed-effect monthly indicator variables and number of spring festival days as covariates. Comparison of changes in healthcare utilization between regions with different SES were conducted using interaction tests [33] [34] [35] . Prior to the conduct of regression analyses, we analyzed data over-dispersion using likelihood ratio tests and confirmed that negative binomial models were more appropriate than Poisson equivalents, as the negative binomial model represents Poisson regression adjusted for excess variation [36] . The negative binomial model equation estimating monthly utilization at provincial level was expressed as follows:

Here, Y it denotes the value of one of the response variables (e.g., monthly outpatient visits at hospitals) included in this study for province i at time t. T t is the time (months) elapsed since the start of the study. T j is the indicator of the post SARS-COV-2 month j (from January to June 2020) . HDI i is the SES category (low (reference group), middle or high) that province i is allocated to. SF t is the number of days of spring festival holiday in month t. Month is the indicator of calendar month of the year with month of January as the reference category. P it is the catchment population for hospitals or THCs in province i at time t . β 0 i represents the model intercept with both a fixed effect and province-level random effects, and β 1 i represents the underlying pre-SARS-COV-2 secular trends with both a fixed effect and province-level random effects. The incidence rate ratio (IRR, model-fitted versus modelexpected/counterfactual level had the pandemic not occurred) of utilization in month j of the post-SARS-COV-2 period for the low HDI group was quantified as exp( β 2 j ). The ratio of IRR, quantified as exp( β 4 j ), compares the IRR of a specific HDI group with that of the low HDI group in month j of the post-SARS-COV-2 period. Tibet and Hubei were excluded from the subnational analysis due to missing data in the pre-and post-outbreak (January and February 2020) period respectively. No missing data points or extreme outliers were found in other provinces after visually inspecting the data by indicator, health facility type and region. Each type of health facility (hospital or THC) and healthcare utilization (inpatient visits or inpatients discharged) was analyzed separately.

We employed fixed-effects models to analyze the impact of the pandemic on healthcare utilization at each subtype of healthcare facility. Comparison between different subtypes of facility were conducted using interaction tests. Potential effects of seasonality, the spring festival and catchment population were also examined. An AR [1] correlation structure and Newey-West Heteroskedasticity and Autocorrelation Consistent (HAC) standard errors were used in mixed-effects and fixed-effect models respectively to accommodate serial autocorrelation in residual errors, per results of the PACF and Cumby-Huizinga general test [37] . Lost health services utilizations were calculated as the differences between the mean fitted values under the full model and the forecasted (counterfactual) model assuming that the pandemic did not occur. To compute 95% confidence intervals around these differences, we used the range from the percentiles 2.5 and 97.5 of 10,0 0 0 sets of predictions per month under each model using the coefficients and multivariate normal distributions of each model [17] . P-values were calculated as the smaller of the proportion of simulated values falling either above or below zero. This value was then multiplied by 2 to represent a 2-sided p-value [17] . We followed STROBE reporting guideline for retrospective observational cohort study [38] . Analyses were conducted in R-version 4.0.2. Statistical significance was set at alpha = 0.05 and all tests were two-sided.

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

The total number of health facilities in the analysis increased Fourteen percent, 82% and 4% of confirmed cases were reported in January, February and from March to June 2020 respectively.

We observed significant decreases in outpatient and inpatient services utilization for health facilities at all levels during the initial pandemic period, in January 2020 ( Table 1 ; Table S1 ; Fig. 1 ). The unadjusted reduction in utilization rates in January 2020 was greater ( p < 0.0 0 01) than the reduction in the average utilizations in January months in the pre-COVID period. However, the most precipitous decline occurred in February 2020 for all regions ( Fig. 1 ;  Fig. S1 ), with a 47.6% (range from 24.8% in THC to 70% in primary care clinic) and 47.7% (range from 26.8% in other health facility to 57.8% in first level hospital) decrease in all-cause visits and inpatient volume, respectively. Rebounds were first noted in March 2020 ( Fig. 1 ; Fig. 2 ; Fig. S1 ), concurrent with a dramatic drop in new SARS-COV-2 cases and the downgrade of the public health emergency response level from Level I to Level II or III in most provinces as virus threat receded.

The adjusted relative change of the monthly rate of national health service utilization by facility type (hospital versus PHCF) and treatment setting (outpatient and emergency department versus inpatient department) are reported in Table 2 . A 51% decrease (IRR: 0.49; 95% CI: 0.48-0.50, p < 0.0 0 01) in health facility visits was observed in February 2020, greater than the 49% reduction (IRR: 0.51; 95% CI: 0.50-0.54, p < 0.0 0 01) in inpatient visits, representing a modest (but statistically significant) 6% reduction (ratio of IRR: 0.94; 95% CI: 0.90-0.98, p = 0.0079) in the incidence rate ratio. The decreases were greater for hospitals than for PHCFs in both health facility visits (59% versus 44%, p < 0.0 0 01) and number of inpatients discharged (51% versus 41%, p < 0.0 0 01). Following the nadir of utilization in February, both health facility visits and inpatient volume showed consistent statistically significant increases from March to June ( Fig.1 ) , rising to 89% (95% CI: 86-92%, p < 0.0 0 01) and 91% (95% CI: 89-93%, p < 0.0 0 01) of the modelbased pre-COVID forecasted (or counterfactual) level had the pandemic not occurred. We estimate a cumulative loss of 1020.5 (95% CI: 951.2-1089.4; P < 0.0 0 01) million or 23.9% (95% CI 22.5-25.2%; P < 0.0 0 01) health facility visits, and 28.9 (95% CI: 26.1-31.6; Table 1 Monthly number of health facility visits and inpatients discharged from January 2016 through June 2020, stratified by health facility type.

Year 2016-20,19 2 Year 2020 P < 0.0 0 01) million or 21.6% (95% CI 19.7-23.4%; P < 0.0 0 01) inpatients nationwide as of June 2020 due to the SARS-CoV-2 pandemic.

The effects were heterogenous across subtypes of hospital and PHCF (Table S2 ). The reductions in the initial pandemic period was comparable between public and private hospitals for both all-cause visits (59% versus 69% down, p = 0.82) and inpatient (51% versus 53% down, p = 0.24) volume. In June, inpatient volume increased to 116% (95% CI: 108-125%; p < 0.0 0 01) of the forecasted level in private hospital but was 14% (95% CI: 12-15%; p < 0.0 0 01) down in public hospitals. Tertiary hospital showed significantly greater reduction in the initial pandemic period and a weaker rebound in the reopening phase than first level hospitals for both outpatient and inpatient volume. Reduction in visits for primary health care in February ranged from 71% (95% CI: 70-72%; p < 0.0 0 01) in primary care clinic to 25% (95% CI: 21-28%; p < 0.0 0 01) in THC, and the reduction in inpatient volume in community health center (56%; 95% CI: 47-63%) was greater ( p < 0.0 0 01) compared with that in THC (39%; 95% CI: 30-38%). Among all indicators, only inpatient volume in private hospital and first level hospital exceeded the forecasted number by June 2020.

Compared with regions in low the HDI category, more developed regions showed significantly greater reduction in healthcare services utilizations in the initial pandemic period and weaker rebound in the reopening phase ( Table 3 ; Table S3 ; Fig. 2 ). In February 2020, a 40% (95% CI: 38-43%, p < 0.0 0 01) model fitted decrease was observed in hospital and THC visits for regions with low HDI, significantly lower ( p < 0.0 0 01 and p < 0.0 0 01) than the 52% (95% Table 2 Comparison of estimated relative change in health care services indicators due to SARS-COV-2 outbreak in hospital and PHCF (January-June 2020) 1 . CI: 50-54%, p < 0.0 0 01) and 55% (95% CI: 53-58%, p < 0.0 0 01) decreases for regions in middle and high HDI groups, respectively. Similarly, the 43% (95% CI: 39-46%, p < 0.0 0 01) decrease in monthly inpatients discharged from hospital and THC in regions of low HDI was lower ( p < 0.0 0 01 and p < 0.0 0 01) than the 51% (95% CI: 48-53%, p < 0.0 0 01) and 53% (95% CI: 49-56%, p < 0.0 0 01) decrease for regions in middle and high HDI group respectively. By June 2020, utilization levels for all indicators were comparable to or higher than expected levels in regions with low HDI, but still significantly lower than expected levels in more developed regions, ranging from 27% (95% CI: 13-39%, p = 0.0 0 07) down for inpatient volume at THC in regions with high HDI to 9% (95% CI: 2-15%, p = 0.0084) down for THC visits in regions with middle HDI (Table S3 ).

The results of our analysis indicate substantial reductions in the monthly volume of health facility visits and inpatient discharges in early 2020, as well as a progressive restoration following the nadir of health service utilization in February 2020, coinciding with the initial SARS-COV-2 outbreak and the subsequent control of the outbreak, respectively. These patterns were apparent for all types of health facilities and throughout the country, even in provinces where very few cases occurred. Utilization rates had not recovered to their pre-SARS-COV-2 levels by June for most indicators, despite no new cases reported in most provinces.

The abrupt reduction in the number of health facility visits was likely attributable to demand, supply and physical accessibility factors associated with the SARS-COV-2 outbreak [13] . First, the lack of knowledge of the virus and the fear of contracting the disease made patients and their family members less likely to visit health facilities [9 , 13 , 15 , 18] . Patients may have chosen to forgo or delay in-person care at health facilities during the pandemic even when the lockdown policies were lifted [13 , 18] . Although there were no additional confirmed cases in most provinces after March 2020, awareness of the ongoing sporadic transmissions mainly caused by imported and sporadic cases elsewhere in the country might have continued to influence health seeking behaviors. Also, there was likely a conversion of many in-person visits to telehealth visits and consultations [39] [40] [41] [42] [43] . Due to the quarantine and social distancing need, there has been a surge in the use of telehealth and virtual care services provided by traditional health facilities or through the platforms of technology companies in China during the pandemic, which has been widely reported [43] [44] [45] [46] . Second, the drop of income and employer-based health insurance coverage due to furlough or loss of job further raised the barriers to accessing healthcare services [47] [48] [49] . The financial and social security of the less-educated, low-income and low-skilled groups, ruralhukou migrant workers in particular, have been disproportionately affected because of the substantially higher unemployment rate among them at this crisis, exacerbating the pre-existing inequalities in access to health services [48 , 50] . Third, restrictions on mobility were implemented extensively across the country. Following Wuhan's lockdown on 23 January, all provinces subsequently initiated the Level I public health emergency response by the end of January. By 4 February, over 95% of cities in each province had implemented transmission control measures either before or after the first case occurred [7 , 8] . About 40% of the cities suspended intra-city public transportation (bus and subway) and 64% of the cities prohibited travel to and from any other cities by any means [7] . Fourth, providers may delay elective health care to reduce the risk of transmitting the virus to either patients or health care workers despite most health facilities remaining open [13 , 18] . In some provinces, health facilities or their outpatient services for non-communicable diseases and non-emergency surgeries were suspended [8 , 18 , 51] . Finally, China's rapid and strict counter-measures not only effectively controlled the spread of SARS-COV-2, but also significantly reduced the incidence and mortality of non-SARS-COV-2 diseases including cardiovascular disease, non-SARS-COV-2 pneumonia, chronic respiratory diseases and injuries, and the healthcare utilization typically associated with these conditions [6] .

The larger reduction in more developed regions and in higher level health facilities also supports the hypothesis that the widespread reduction in health care utilization in China was attributable to the SARS-COV-2 epidemic. Due to the low number of confirmed cases and the successful mitigation of the transmission, more provinces in the lower HDI group downgraded their public health emergency response level from level I to Level II or III in February [3 , 7 , 52 , 53] , resuming intercity travel and intracity public transport. Provinces with lower HDI are more likely to have greater rural populations. Given media coverage about SARS-COV-2 prevention that focused predominantly on large urban cities, rural residents reported less positive attitudes toward the effectiveness of performing preventive behaviors; as such, rural residents were less likely to engage in behaviors such as avoiding gathering, staying home as much as possible and avoiding public transportation [10] . Moreover, the inter-city and inter-provincial travel ban was more likely to prevent patients from rural and less developed regions from seeking health care at health facilities in more developed regions. The urge to implement hierarchical medical care to avoid crowds in tertiary hospitals during the epidemic may also partially explain the greater losses of visits in tertiary hospitals.

We found that nearly all healthcare utilization measures had not recovered to the pre-SARS-COV-2 levels by June 2020, despite very few new SARS-COV-2 cases since April and the removal of travel restrictions across the country in March 2020 [3 , 54] . These reductions and stagnations in prevention and treatment will likely have significant collateral effects on population health that greatly exceed the direct health effects from the infection [9 , 16 , 17 , 26 , 55] . In China, only a small portion of the unemployed receive unemployment insurance. Despite some flexibility (indicated by the government in May) in offering one-time subsidies to the unemployed not enrolled in the unemployment insurance system, the new policy applied only to a very limited number of exceptional cases [48] . Crippling losses in revenue owing to the precipitous decline of patient visits threaten the viability of a substantial number of healthcare facilities and providers, especially private primary care clinics and village practices that were already financially vulnerable [49] . In this context, our findings call for renewed efforts to educate patients about safety protocols, as well as targeted interventions to restore community trust and confidence in health facilities [13 , 14] . Also, our findings suggest that flexible and broad-based compensation programs may be needed to alleviate financial barriers to accessing healthcare services among vulnerable populations [4 8 , 4 9 ] . The findings also suggest the need to adopt innovative payment models such as a capitation model that decouples compensation from the volume of services provided; such approaches could help stabilize facility and provider finances compared to the current fee-for-service system, especially in the extreme circumstance of a pandemic [49] .

To our knowledge, this study is the first to quantify the effect of the SARS-COV-2 outbreak on longitudinal trends in health services utilization in China, where the disease was first reported. Our analyses are comprehensive, with detailed monthly data spanning multiple years before, during, and after the outbreak, covering the entire population, and using robust statistical methods. These analyses show the value of routine health information system (RHIS) data for timely and effectively tracking the performance of health system in low-and middle-income countries [17 , 56] .

Our results are subject to several important limitations. First, the possibility of data collection errors and delay in reporting can-not be ruled out even though routine data were extracted and reported using standardized procedures. If so, the use of the aggregate RHIS data may result in some bias in the estimates. Second, given the nature of the aggregated data, we were not able to explore the cross-province heterogeneity by health facility type, nor were examinations by patient-level demographic variables (sex and age) or disease type feasible. Third, we were not able to estimate the utilization losses in Hubei in the first two months of the outbreak, January and February 2020, due to non-reporting of the outcomes. However, the exclusion of Hubei in the analysis at national and subnational level would most likely not significantly affect the results because that patient volume in Hubei in a typical month accounted for less than 4% of the national total and the recovery patterns for both outpatient and inpatient volume in Hubei after February were comparable to those in other regions that had implemented similar public health emergency response (Fig. S1) . Also, tests for interaction have notably lower power than tests for main effects [57] ; thus analyses of differential trends between groups that lack statistical significance may represent false negative findings. Last, our assessment of health services utilization could not capture trends in the use of telehealth visits and consultations, nor were we able to quantify potential changes in the quality of care before, during and after the SARS-COV-2 outbreak.

In conclusion, our study shows that outpatient and inpatient volume in healthcare facilities at all levels significantly declined in conjunction with the SARS-COV-2 outbreak, and most utilization measures have not returned to their pre-outbreak levels as of June 2020, despite the fact that China rapidly and effectively controlled the pandemic. Sustained monitoring and targeted interventions that restore public trust and confidence in healthcare facilities and protect providers' financial risk are needed to mitigate the collateral disruption on healthcare services. Future research is needed to establish which vulnerable patient groups and which disease-specific healthcare services were most affected by the pandemic, in order to continue to inform strategies and policies aimed at mitigating the impact of potential future pandemics.

HX, YG and JMU conceived the research question and developed the protocol. YG and JMU oversaw study implementation. HX was responsible for the data curation and analysis and the writing of the first draft of the manuscript. XD, BHW, FL, OA, YG and JMU assisted with development of the study design, data analysis, data interpretation, and critical review of the manuscript. HX, YG and FL have assessed and verified the underlying data and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript.

The data that supports the findings of this research are publicly available from the Center for Health Statistics and Information: http://www.nhc.gov.cn/mohwsbwstjxxzx/s2906/new _ list.shtml .

The National Natural Science Foundation of China (No. 72042014 ).

The authors declare no competing interests related to the study.

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.lanwpc.2021.100122 .

The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak

An investigation of transmission control measures during the first 50 days of the COVID-19 epidemic in China

Response to the COVID-19 epidemic: the Chinese experience and implications for other countries

National Public Health Emergency Response Protocol

Investigating public health emergency response information system initiatives in China

Do lockdowns bring about additional mortality benefits or costs? Evidence based on death records from 300 million chinese people

The effect of human mobility and control measures on the COVID-19 epidemic in China

The public health response to the COVID-19 outbreak in mainland China: a narrative review

Effect of Ebola virus disease on maternal and child health services in Guinea: a retrospective observational cohort study

Differences in preventive behaviors of COVID-19 between urban and rural residents: lessons learned from a cross-sectional study in China

The impact of the COVID-19 epidemic on the utilization of emergency dental services

Collateral effect of Covid-19 on stroke evaluation in the United States

Larson DB . Variables influencing radiology volume recovery during the next phase of the coronavirus disease 2019 (COVID-19) pandemic

COVID-19 pandemic impact on decreased imaging utilization: a single institutional experience

Impact of the COVID-19 epidemic on stroke care and potential solutions. Stroke

COVID-19 exacerbating inequalities in the US

The 2014-2015 Ebola virus disease outbreak and primary healthcare delivery in Liberia: time-series analyses for 2010-2016

Indirect implications of COVID-19 prevention strategies on non-communicable diseases: an opinion paper of the european society of hypertension working group on hypertension and cardiovascular risk assessment in subjects living in or emigrating from low resource settings

The impact of the SARS epidemic on the utilization of medical services: SARS and the fear of SARS

Impact of SARS on healthcare utilization by disease categories: implications for delivery of healthcare services

Urban-rural disparities in mental health problems related to COVID-19 in China

Hospital admissions for acute myocardial infarction before and after lockdown according to regional prevalence of COVID-19 and patient profile in France: a registry study

The Covid-19 Pandemic and the Incidence of Acute Myocardial Infarction

Incidence of acute myocardial infarction-related cardiogenic shock during corona virus disease 19 (COVID-19) pandemic

Impact of lockdown on bed occupancy rate in a referral hospital during the COVID-19 pandemic in northeast Brazil

Effect of the Ebola-virus-disease epidemic on malaria case management in Guinea, 2014: a cross-sectional survey of health facilities

Paediatric care in relation to the 2014-2015 Ebola outbreak and general reporting of deaths in Sierra Leone

Reduction of hospitalizations for myocardial infarction in Italy in the COVID-19 era

Association of the coronavirus disease 2019 (COVID-19) outbreak with enrollment in cancer clinical trials

Ministry of Health's Notice on Establishing the Direcct Reporting Health Information System (Chinese)

Reflection on modern methods: a common error in the segmented regression parameterization of interrupted time-series analyses

Interrupted time series regression for the evaluation of public health interventions: a tutorial

Increased inequalities in health resource and access to health care in Rural China

Socioeconomic inequalities of outpatient and inpatient service utilization in China: personal and regional perspectives

Socioeconomic inequities in health care utilization in China

Testing approaches for overdispersion in poisson regression versus the generalized poisson model

Testing the autocorrelation structure of disturbances in ordinary least squares and instrumental variables regressions

The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies

COVID-19 transforms health care through telemedicine: evidence from the field

Telemedicine during the COVID-19 pandemic: experiences from Western China

Using telemedicine during the COVID-19 pandemic

Rethinking online mental health services in China during the COVID-19 epidemic

Online mental health services in China during the COVID-19 outbreak

Use of telemedicine and virtual care for remote treatment in response to COVID-19 pandemic

How telemedicine integrated into China's anti-COVID-19 strategies: case from a National Referral Center

Covid-19 gives boost to China's telemedicine industry: the wall street journal

The impact of the COVID-19 pandemic on firms: a survey in Guangdong Province

Unequal pain: a sketch of the impact of the Covid-19 pandemic on migrants' employment in China

Covid-19 -implications for the health care system

Who loses income during the COVID-19 outbreak? Evidence from China

COVID-19 infection epidemic: the medical management strategies in Heilongjiang Province

Five more Chinese regions lower emergency response level as virus threat recedes 2020

Three more Chinese regions lower emergency response level as virus threat recedes 2020

China's successful control of COVID-19

Impact of COVID on college student mental health and wellness

Using routine health information systems for well-designed health evaluations in low-and middle-income countries

Interaction revisited: the difference between two estimates