Background

The pandemic of coronavirus disease 2019 (COVID-19) has been the latest threat to global health since the first case was reported in China in late December 2019 [1]. Thailand, an upper middle-income country located in South-East Asia and not far from China, had rather highly effective response at the start of the COVID-19 pandemic, which began in March 2020, and experienced relatively few COVID-19 cases compared to other countries during the start of pandemic [1]. This pattern was contributed mainly due to several non-pharmaceutical measures (NPI) implemented during the beginning of COVID-19 pandemic including isolation and quarantine of detected cases, use of face masks and hand hygiene, and social distancing as well as full-scale national lockdown from March 26 to May 3, 2020 (i.e., night curfews, travel ban, and closures of school and all public spaces) [2]. As out-patient department (OPD) is a crucial component for improved service delivery and efficiency enhancement specially in context of healthcare systems in low- and middle-income countries (LMICs) [3, 4], the impact of COVID-19 pandemic and its measures on utilisation of same was expected to be substantial. For instance, studies in Kenya and China showed that the number of OPD visits had dropped by approximately 35–60% during the early period of pandemic [5, 6].

Thailand has three main health care schemes including the Civil Servant Medical Benefit Scheme (CSMBS), Social Security Scheme (SSS), and Universal Coverage Scheme (UCS) [7]. The UCS, established since 2002, is being managed by the National Health Security Office (NHSO) [8] and covered approximately 80% of the population in Thailand or 48 million people in 2017. Therefore, analysing the pattern of OPD utilisation during the COVID-19 pandemic using the UCS’ national database from NHSO could aid with the planning for future and predict the readiness on pandemic preparedness response. This study aims to estimate the pattern of OPD utilisation during COVID-19 in Thailand, which included overall utilisation and utilisation in each diagnostic group, age group, and health region.

Methods

This study was a secondary data analysis of outpatient data from patients covered under the UCS in Thailand. Data were obtained from the NHSO over a 4-year period between January 2017 and December 2020 which was separated into 3 periods: 1) the pre-pandemic period (January 2017 to March 2020); 2) the lockdown (April – May 2020); and 3) the post-lockdown period (June – December 2020). The number of patients for each month were aggregated into 22 diagnostic groups (based on the 10th revision International Classification of Diseases [ICD-10]) (Additional file 1) [9]; 7 age groups (10-year age bands from 0 to 60 and older than 60 years); and 13 health regions throughout Thailand [10]. The OPD utilisation was represented by the number of OPD visits divided by number of UCS beneficiaries in each group. The study was approved by the ethics committee of Institute for the Development of Human Research Protections.

Statistical Analysis

An interrupted time series (ITS) analysis was performed to estimate the impact of COVID-19 on pattern of OPD utilisation in Thailand. The impact was modelled using a segmented Quasi-Poisson regression with pre-pandemic trends (January 2017 – March 2020) and indicator variables for lockdown (April–June 2020) and post-lockdown periods (July – December 2020) as follows:

$${Y}_{t}={\beta }_{0}+{\beta }_{1}T+{\beta }_{2}{X}_{t}+{\beta }_{3}T{X}_{t}+\epsilon$$

where \({X}_{t}\) represented the dummy variable indicating the pre-pandemic (coded 0), lockdown (coded 1) or post-lockdown periods (coded 2), T was time in months from January 2017, and \({Y}_{t}\) was the OPD utilisation at time t. In the model, \({\beta }_{0}\) estimates the baseline level at T = 0, \({\beta }_{1}\) estimates the change in outcome associated with a time unit increase, \({\beta }_{2}\) estimates the level change following the intervention, and \({\beta }_{3}\) is the slope change following the interaction between time and intervention.

The rate ratios (RRs) and their 95% confidence interval (95% CI) of OPD utilisation during lockdown and post-lockdown periods compared to OPD utilisation in the pre-pandemic period were calculated. The RRs for overall OPD utilisation as well as for each diagnostic group, age groups, health regions, and provinces were explored. All statistical analyses were performed using R version 4.1.3 (R Foundation for Statistical Computing, Vienna, Austria) [11]. An ITS was conducted using The packages lmtest [12], vcd [13], Epi [14], tsModel [15], and splines. A p-value of 0.05 or lower was considered statistically significant (i.e. significant change).

Results

A total of 845,344,946 OPD visits was observed during January 2017 to December 2020. There was an increasing trend in monthly OPD visits per 100 UCS beneficiaries across the pre-pandemic period (Fig. 1).

Fig. 1
figure 1

Number of monthly OPD visits per 100 UCS beneficiaries from January 2017 to December 2020. Note: Black dots = average number of monthly OPD visits per 100 UCS beneficiaries from January 2017 to December 2020; Red solid lines = seasonally adjusted trend of monthly OPD visits per 100 UCS beneficiaries based on the regression model; Blue dashed lines = de-seasonalized trend of monthly OPD visits per 100 UCS beneficiaries

Full size image

The seasonal patterns were also observed with the highest and lowest number of OPD visits occurring in January and September, respectively. The number of OPD visits decreased by 28% (RR 0.718, 95% CI: 0.631–0.819) and 11% (RR 0.890, 95% CI: 0.811–0.977) during the COVID-19 lockdown and post-lockdown periods, accordingly, when compared to the pre-pandemic period.

The relative change in OPD visits during and after COVID-19 lockdown in each diagnostic group is shown in Table 1 and Additional file 2.

Table 1 Relative change in OPD visits during and after COVID-19 lockdown, stratified by diagnostic groups
Full size table

There was a significant decline in the number of OPD visits for all diagnostic groups during the lockdown period, except for ICD-10 codes of V01–Y98 (external causes of morbidity and mortality such as transport accidents or assault) and O00–O99 (pregnancy, childbirth, and the puerperium). The largest drop in OPD utilisation was observed in diseases of the respiratory system (ICD-10: J00–J99) followed by certain conditions originating in the perinatal period (P090–P96) with RRs of 0.411 (95% CI: 0.320–0.527) and 0.489 (95% CI: 0.322–0.743), respectively. A significant decline in the number of OPD visits due to respiratory diseases after the lockdown remained until the end of 2020 (RR 0.637, 95% CI: 0.551–0.737). Meanwhile, the numbers of OPD visits for endocrine, nutritional, and metabolic diseases (E00–E90), diseases of the circulatory system (I00–I99), and diseases of the genitourinary system (N00–N99) during the post-lockdown period were not significantly different from the pre-pandemic period.

During the lockdown period, the number of OPD visits in all age groups was significantly decreased (Table 2 and Additional file 3).

Table 2 Relative change in OPD visits during and after COVID-19 lockdown, stratified by age groups
Full size table

The younger the age groups, the higher the impact was observed. For example, a drop by almost half was detected in 0–10 age groups with a RR of 0.545 (95% CI: 0.419–0.709), whereas a drop of only 20% (RR 0.801, 95% CI: 0.696–0.922) was observed in a group of 60 years and above. The number of OPD visits after the lockdown in the age group of older than 30 years was similar to their pre-pandemic levels while the age group of 30 years or younger reported significantly lower than their pre-pandemic levels.

Table 3 and Additional file 4 demonstrate the relative change in OPD visits during and after COVID-19 lockdown in 13 health regions across the country [10]. Thailand has 77 provinces which were grouped into 13 health regions. Each health region has between four to eight provinces with one exception for health region 13 which includes only 1 province, Bangkok, the capital (and biggest province) of Thailand.

Table 3 Relative change in OPD visits during and after COVID-19 lockdown, stratified by health regions
Full size table

The number of monthly OPD visits in health region 13 (Bangkok, the capital of Thailand) was approximately 10 visits per 100 UCS beneficiaries when compared to 40 visits in other health regions. A significant decrease in the number of in OPD visits during lockdown period was observed in all health regions, except in health region 13 (Bangkok). The strongest impact of the lockdown was noted in health region 11 (southern provinces), specifically the number of OPD visits fell by 10–20% after the lockdown when compared to the pre-pandemic level. There was no significant difference in the number of OPD visits after the lockdown in health regions 1 and 2 (northern provinces), 9 and 10 (northeastern provinces), and 13 (Bangkok).

Discussion

This analysis of a big data on OPD visits in Thailand showed an impact of the COVID-19 lockdown on ambulatory service utilisation. The minimal impact can be seen for planned visits as the number of visits among non-communicable diseases and in elderly were only trivially dropped. On the contrary, the higher impact was observed for diseases of respiratory system, which can be corelated to a decreased incidence due to social distancing measures. The findings reassured that the services for those patients who have been planned in advanced were only slightly affected during the lockdown period. After the lockdown, the number of OPD visits returned to their pre-pandemic levels in most diagnostic groups, age groups, and health regions.

The minimal impact of the COVID-19 lockdown on OPD visits among adults and elderly was observed in our study with a decline rate of approximately 20–25% whereas a declining rate in children and adolescent (0–10 and 10–20 years) was around 45%. Moreover, the observed impact of COVID-19 lockdown appeared to be similar across health regions with one exception of health region 13 which is Bangkok, the capital of Thailand. This variation might be due to the differences between the capital and other parts of the country such as the highest population density for the capital and the health service structure of Bangkok [16]. The varying degree of declines for each service was also reported in other countries including Ethiopia, Haiti, Ghana, Lao People’s Democratic Republic, Mexico, Nepal, South Africa, Chile, and South Korea [17]. Moreover, the number of OPD visits in most diagnostic groups, age groups, and health regions returned to the pre-pandemic levels once the lockdown was lifted. A slight impact on the planned OPD visits might be due to a relatively small outbreak of COVID-19 throughout 2020. For instance, a total of 4,862 COVID-19 cases and 62 deaths were reported in 2020 [18]. The estimated excess all-cause deaths associated with COVID-19 were -22 (95% CI: -34 to -11) in 2020 compared with 44 (95% CI: 26–62) in 2021–2022 [19].

The COVID-19 control measures had a positive effect on the incidence of respiratory diseases especially during the lockdown period. This trend was also observed in Singapore and the United States [20, 21]. However, the difference in magnitude might be due to variation in stringency index in each country [22]. In our study, the effect of COVID-19 control measures on respiratory diseases was persisted long after the lockdown, conversely, a re-emergence of respiratory viral infections was found in Singapore after 13 weeks of the lockdown [21]. A high proportion of population wearing mask and washing their hands in Thailand despite lockdown easing [23] and a higher population density in Singapore [24] might explain the results.

To date, there are a few studies evaluating the impact of COVID-19 on OPD utilisation nationally [5, 6, 17, 25]. We analysed more than 800 million OPD records over 4-year period in this study, which covered ~ 80% of the population in Thailand. Knowing the OPD utilisation pattern during the COVID-19 pandemic could be used to answer health policy-related questions. For example, a positive effect of COVID-19 control measures on respiratory diseases could support a continuation of mask wearing in public spaces.

There are limitations to our study since we used an aggregated data. First, demographic and clinical characteristics information, except age group and health region, were not available for confounding adjustment. Second, diagnoses in each ICD-10 diagnostic group could not be specified. We also could not explore adverse health outcomes (i.e., hospitalisation rate, mortality rate) after a reduction in OPD utilisation due to lack of relevant data. Finally, the current study only focused on the first year of the pandemic; therefore, future research should explore and compare the OPD utilisation patterns in subsequent years including detailed types of visits (e.g., emergency cases) to support further planning for our healthcare system.

In conclusion, we found that the OPD utilisation rate in Thailand decreased during the COVID-19 lockdown with a varying impact on different service and returned to pre-pandemic levels after the lockdown. The results from this study suggest that healthcare providers might have considered both policy and pandemic situation when implementing control measures for each health sector. Further studies on healthcare providers’ attitude and behaviour toward the lockdown policy could provide a better understanding of which service sector should be prioritised and strengthen the healthcare system for future pandemics.