Pandemic trends in health care use: From the hospital bed to self-care with COVID-19.

AIM: To explore whether the acute 30-day burden of COVID-19 on health care use has changed from February 2020 to February 2022.
METHODS: In all Norwegians (N = 493 520) who tested positive for SARS-CoV-2 in four pandemic waves (February 26th, 2020 -February 16th, 2021 (1st wave dominated by the Wuhan strain), February 17th-July 10th, 2021 (2nd wave dominated by the Alpha variant), July 11th-December 27th, 2021 (3rd wave dominated by the Delta variant), and December 28th, 2021 -January 14th, 2022 (4th wave dominated by the Omicron variant)), we studied the age- and sex-specific share of patients (by age groups 1-19, 20-67, and 68 or more) who had: 1) Relied on self-care, 2) used outpatient care (visiting general practitioners or emergency ward for COVID-19), and 3) used inpatient care (hospitalized ≥24 hours with COVID-19).
RESULTS: We find a remarkable decline in the use of health care services among COVID-19 patients for all age/sex groups throughout the pandemic. From 83% [95%CI = 83%-84%] visiting outpatient care in the first wave, to 80% [81%-81%], 69% [69%-69%], and 59% [59%-59%] in the second, third, and fourth wave. Similarly, from 4.9% [95%CI = 4.7%-5.0%] visiting inpatient care in the first wave, to 3.6% [3.4%-3.7%], 1.4% [1.3%-1.4%], and 0.5% [0.4%-0.5%]. Of persons testing positive for SARS-CoV-2, 41% [41%-41%] relied on self-care in the 30 days after testing positive in the fourth wave, compared to 16% [15%-16%] in the first wave.
CONCLUSION: From 2020 to 2022, the use of COVID-19 related outpatient care services decreased with 29%, whereas the use of COVID-19 related inpatient care services decreased with 80%.

Introduction

The clinical course and mortality of COVID-19 among hospitalized individuals have been well described, especially for the start of the pandemic (spring 2020) [1-4]. Less is known about COVID-19’s total use of health care services, including both inpatient and outpatient care both during the 1st wave, but especially during the 2nd, 3rd, and 4th waves striking spring and fall 2021 in Norway. To date, studies of outpatient care use during the pandemic have focused on structural changes in its delivery, such as telemedicine visits vs. office-based visits [5], whereas studies of the use of outpatient care services (such as general practitioners and emergency wards) are scarce. The use of health care services may be hypothesized to have changed throughout the pandemic, starting with limited test availability and many persons in risk groups being hospitalized and eventually dying, to mass testing, mass vaccination of persons at risk and an increasing herd immunity. In the latest wave, the health care services may be hypothesized to be better trained in how to manage severely ill patients, leading to declining inpatient care treatment and fewer fatal outcomes in the end than in the beginning of the pandemic.

A timely and correct up- and downscaling of health services (and lockdown measures) depend on our understanding of the pathways patients take through the health system, including the peaks and total demand of health care services following an individual’s positive test. Studying the impact on both inpatient and outpatient care in its early waves, can provide valuable insight into health service needs in later stages, and contribute to the knowledge base that can increase our resilience against future pandemics.

We have access to high quality and very recent register data covering the first three pandemic waves, as well as the first weeks of the fourth wave. Thus, in this paper we aimed to explore the age- and sex-specific acute burden of COVID-19 on the health care services in four waves of the pandemic in Norway through a national descriptive cohort study design using registry data. Because of previously reported differences in vaccination status throughout the pandemic [6], differences in disease severity by SARS-CoV-2 variant [7, 8] and strata of age and sex [9], we hypothesized that we would also see differences in the 30-day pattern of healthcare use for men and women, girls and boys, the working age population, and the elderly in the different waves of the pandemic.

Materials and methods

The BeredtC19-register is an emergency preparedness register aiming to provide rapid knowledge about the pandemic, including impacts of measures to limit the spread of the virus on health and utilization of health care services [10]. BeredtC19 compiles daily updated individual-level data from several registers. It includes the Norwegian Surveillance System for Communicable Diseases (MSIS) (all testing for COVID-19), the Norwegian Patient Register (NPR) (all electronic patient records from all hospitals in Norway), and the Norway Control and Payment of Health Reimbursement (KUHR) Database (all consultations with all general practitioners and emergency outpatient health care), as well as the National Population Register (age, sex, country of birth, date of death). Thus, the register includes all polymerase chain reaction (PCR) tests for SARS-CoV-2 in Norway with date of testing and test result, reported from all laboratories in Norway and all electronic patient records from primary care as well as outpatient and inpatient specialist care. The establishment of an emergency preparedness register forms part of the legally mandated responsibilities of The Norwegian Institute of Public Health (NIPH) during epidemics. Overall, data from Norwegian health registers have been demonstrated to be of high quality with high validity and reliability, and together they can provide a complete picture of patterns of healthcare use [11-13]. Medical recording to the National registries is mandated by law in Norway, ensuring no missing data in our study. The Ethics Committee of South-East Norway confirmed (June 4th, 2020, #153204) that external ethical board review was not required.

Population

Our population included every Norwegian resident who tested positive for the SARS-CoV-2 by a PCR-test from February 26th, 2020, to January 14th, 2022. The date with the first record of a confirmed test was coded as being the start of the individual’s health care pathway. Patients with negative PCR-tests, as well as patients with suspected COVID-19 and without positive PCR-tests were excluded. For persons testing positive multiple times we included a wash-out period of 90 days [14]. We divided our population into mutually exclusive age and sex groups, i.e., girls and boys, men, and women by the following age categories: 1–19 (children and adolescents), 20–67 (working age population) and 68 years or older (elderly), as COVID-19 has hit differently among different age groups and sexes [9].

Outcomes

Our outcomes were defined as follows:

Self-care: No registered health care use (outpatient or inpatient) within 30 days of a positive test.

Outpatient care: Outpatient care use with International Classification of Primary Care (ICPC-2) code R991 or R992 (COVID-19) (general practitioners or emergency wards).

Inpatient care: Hospital-based inpatient specialist care with International Classification of Disease (ICD-10) code U071 (confirmed COVID-19) or U072 (suspected COVID-19).

Death: Death independent of cause but occurring within 30 days after the positive test.

The dates of all outcomes were sorted relative to the date of the positive PCR-test, with the test date being coded as day 0 and the outcomes occurring on day -2 to day 30. We chose a 30-days-timeframe because a death after COVID-19 was classified as covid-related if it occurred within 30 days after testing positive in official statistics [15], and because it coincides with what is commonly referred to as the acute phase of SARS-CoV-2 [16]. Thus, we regarded people who were still alive after 30 days as recovered. When combined, and sorted chronologically on dates of occurrence, these data provided a comprehensive picture of the peak and total use of outpatient and inpatient care, as well as COVID-19-related health care pathways in the acute phase. To account for the fact that many PCR-tests were prescribed by the health care services, we ran analyzes both including and excluding health care use related to the testing.

Study setting

The COVID-19 epidemic in Norway has, since the first registered case on February 26th, 2020, been dominated by four different lineages (Fig 1). In this article we define four waves of the pandemic based on the dominating virus lineages. The 1st wave of transmission (February 26th, 2020 –February 16th, 2021) was characterized by low test availability (only available for health personnel, elderly, and persons at risk) until the summer of 2020, and was primarily dominated by the Wuhan strain of the virus as well as some minor cases of other lineages such as Beta and Gamma. The 2nd wave of transmission (February 17th, 2021 –July 10th, 2021) was characterized by wide testing criteria and free testing, start of vaccination, and was dominated by the Alpha variant. The 3rd wave of transmission (July 11th, 2021 –December 27th, 2021) was characterized by continued wide testing criteria and free testing, most people vaccinated with at least one dose and was dominated by the Delta variant. We also included data from the 4th wave of transmission (December 28th, 2021 –January 14th, 2022), which was dominated by the Omicron variant and saw start of booster vaccination [17-19] (Fig 1).

Fig 1

Daily covid-19 cases by variant and share of population vaccinated.

The top panel shows the number of persons with a positive PCR-test each day calculated with a 7-day moving average. The solid line in the top panel shows the total numbers of persons infected, and colors represent the share of screened or sequenced tests with each variant the given day. The bottom panel shows the cumulative share of the included population having received at least one dose of SARS-CoV-2 vaccine on the given day.

Statistical analyses

In this national descriptive cohort study using registry data, we first assessed descriptive statistics of our study population by the four different waves or periods of the pandemic.

Second, we studied the peak and total use of outpatient and inpatient care during the -2 to +30 days following positive test, for each of the waves. To explore the wave-wise peak use, we estimated day-by-day proportions in need of outpatient or inpatient care at least once during day -2 to +30. For persons with at least one outpatient care visit we included a maximum of one visit per day, and for persons with at least one inpatient care visit, we coded all the days spent in hospital as hospital bed-days. To explore the total use in each wave we estimated the cumulative proportions seeking outpatient or inpatient care at least once during day -2 to +30. We repeated these analyses, excluding day –2 to 0 to account for the impact of testing at the peak and total use of health care. Among persons seeking outpatient or inpatient care at least once, we estimated the mean number of outpatient visits or the mean number of hospital bed days in each wave, by age and gender.

Finally, based on the total health care use observed during the -2 to +30 days following positive test, we divided the study population into three different mutually exclusive major patient pathways. Each of the patient pathways represented a different acute burden of disease on the health care systems: 1) persons relying on self-care with no health care use, 2) patients who had contact with outpatient care (GP and/or emergency ward) only, and 3) patients who had contact with inpatient care, with or without additional need for outpatient care. For each age-/sex-group and for each of their pathways in each of their pandemic wave, we estimated the proportions having the different pathways and calculated 95% confidence intervals. To get an overarching picture of the major patient flows, we visualized the timing of care for these different pathways in alluvial diagrams. We also estimated the whole-sample- and pathway-specific mortality as proportions with 95% confidence intervals. 95% confidence intervals were calculated as where μ is the mean, σ is the standard deviation and n is the population size. All analyses were run using STATA SE v.16.

Results

We identified 493 520 persons with at least one positive PCR-test for SARS-CoV-2 in the total tested population of 3 625 617 persons between February 26th, 2020, and January 14th, 2022. The total number of tests throughout the pandemic was 10 520 144. The percentages of positive tests among all tests in the 1st, 2nd, 3rd, and 4th waves were 1.9%, 2.2%, 8.0%, and 27.2%, respectively. Table 1 shows that the average age of persons testing positive decreased from the 1st to the 4th wave. It also shows that the percentage of women among those testing positive increased from the first two waves to the last two waves (Table 1). The proportions dying within 30 days after positive test decreased throughout the pandemic, from 11.3% to 1.8% for persons above 68 years (S1 Table). The 30-day mortality was low across all pandemic waves for persons aged under 67 years (S1 Table).

Table 1

Descriptive characteristics of persons testing positive for SARS-CoV-2 in each of four pandemic waves in Norway, 2020–2022.

		1st wave	2nd wave	3rd wave	4th wave
		Feb 26th ‘20—Feb 16th ‘21	Feb 17th ‘21 –Jul 10th ‘21	Jul 11th ‘21—Dec 27th ‘21	Dec 28th ‘21 –Jan 14th ‘22
		N = 64 254	N = 64 181	N = 244 517	N = 120 568
Characteristics of the whole sample
Age, mean (SD)		37.0 (19.5)	30.2 (17.9)	28.7 (19.5)	28.4 (17.5)
Women, N (%)		30533 (47.5)	30107 (46.9)	120531 (49.3)	59779 (49.6)
Outcomes
	Self-care (%)	9941 (15.5)	12039 (18.8)	75,242 (30.8)	49,216 (40.8)
	Outpatient care (%)	53549 (83.3)	51850 (80.8)	168200 (68.8)	71107 (59.0)
	Inpatient care (%)	3129 (4.9)	2294 (3.6)	3329 (1.4)	570 (0.5)
	All-cause mortality (%)	590 (0.9)	146 (0.2)	571 (0.2)	52 (0.0)
Share vaccinated by start of the wave (min. 1 dose)		0%	5%	57%	79%
Characteristics of the studied strata
Children and adolescents (1–19 years)
	Girls, N (%)	6119 (9.5)	10601 (16.5)	49988 (20.4)	21558 (17.9)
	Boys, N (%)	6521 (10.1)	11051 (17.2)	51980 (21.3)	22049 (18.3)
Adults in working age (20–67 years)
	Women, N (%)	21949 (34.2)	18702 (29.1)	65883 (27.0)	36996 (30.7)
	Men, N (%)	25031 (39.0)	22133 (34.5)	67387 (27.6)	37553 (31.2)
Elderly (≥68 years)
	Women, N (%)	2449 (3.8)	792 (1.2)	4654 (1.9)	1225 (1.0)
	Men, N (%)	2185 (3.4)	902 (1.4)	4625 (1.9)	1187 (1.0)

Peak and total use of outpatient care

In the early phases of the pandemic, patients sought outpatient care frequently. 83% of all who tested positive sought outpatient care within 30 days in the 1st wave. This share decreased only marginally for the 2nd wave (81%), and decreased further to 69% in the 3rd wave, and to 59% in the 4th wave. During the 1st wave, 40% of all patients sought outpatient care on day 0 (including day -1 and -2). This day-0-proportion decreased to ~30% in the 2nd and 3rd wave, and further to 25% in the 4th wave (Fig 2). Furthermore, during the 1st wave, patients continued to visit outpatient care for a longer time compared to the other waves, with 2.4% still visiting outpatient care on day 20, compared to 1.5% in the 2nd wave, 0.6% in the 3rd wave, and 0.3% in the 4th wave (Fig 2). S2 Fig shows that middle-aged and elderly persons continued to seek outpatient care for a longer time than persons aged 1–19 years. For all age groups, the mean number of visits (among those seeking outpatient care) steadily decreased for each wave of the pandemic, from ranging between 2–4 visits per patient, to 1–2 per patient in the 4th wave (Fig 3). When we excluded outpatient care use in relation to testing, results showed no change in the cumulative share from the 1st wave (76%) to the 2nd wave (75%), a slight decrease in the 3rd wave (58%) and continued decrease in the 4th wave (45%) (S3 Fig).

Fig 2

Shares visiting outpatient care.

Day by day and cumulative shares seeking outpatient care (GP or emergency ward) from the day of testing positive (day 0) to the 30th day after positive test, for four waves/periods of the pandemic: February 26th, 2020 –February 16th, 2021 (1st period), February 17th, 2021 –July 10th, 2021 (2nd period), July 11th, 2021 –December 27th, 2021 (3rd period), and December 28th, 2021 –January 14th, 2022 (4th period).

Fig 3

Number of outpatient care visits.

The mean (95% confidence intervals) number of visits in outpatient care, for women and men in different age groups, four waves/periods of the pandemic: February 26th, 2020 –February 16th, 2021 (1st period), February 17th, 2021 –July 10th, 2021 (2nd period), July 11th, 2021 –December 27th, 2021 (3rd period), and December 28th, 2021 –January 14th, 2022 (4th period), among persons having at least one visit in outpatient care.

Peak and total use of inpatient care

We observed an even more considerable shift in the total use of inpatient care throughout the pandemic. From 4.9% of all who tested positive being hospitalized at least once during the 1st wave, compared to 3.6% during the 2nd wave, 1.4% in the 3rd wave, and 0.5% in the 4th wave (Fig 4). For the three first waves, the share seeking inpatient health care peaked between the 8th and 11th day following a positive test (Fig 4). For the fourth wave, the peak was observed on the first day. S4 Fig shows that a minor proportion of children and adolescents, and a considerable proportion of elderly sought inpatient care. Along this line, the mean number of bed-days was also higher for the middle-aged and elderly, than it was for children (Fig 5). Generally, women tended to have a lower number of bed-days than men (Fig 5). For men and women in their working age (20–67 years), and for men aged ≥68 years the mean number of bed-days decreased from the first to the fourth wave (Fig 5). No statistically significant decrease was observed for children or for women aged 68 or more (Fig 5). As expected, we see little changes in the cumulative share of persons seeking inpatient care when excluding days –2 to 0, i.e., during the period in which any health care use could be assumed to be related to testing (S5 Fig).

Fig 4

Shares seeking inpatient care.

Day by day and cumulative shares seeking inpatient care from the day of testing positive (day 0) to the 30th day after positive test, during four waves/periods of the pandemic: February 26th, 2020 –February 16th, 2021 (1st period), February 17th, 2021 –July 10th, 2021 (2nd period), July 11th, 2021 –December 27th, 2021 (3rd period), and December 28th, 2021 –January 14th, 2022 (4th period).

Fig 5

Number of bed-days spent in inpatient care.

The mean (95% confidence intervals) number of bed-days spent in inpatient care, for women and men in different age groups, during four waves/periods of the pandemic: February 26th, 2020 –February 16th, 2021 (1st period), February 17th, 2021 –July 10th, 2021 (2nd period), July 11th, 2021 –December 27th, 2021 (3rd period), and December 28th, 2021 –January 14th, 2022 (4th period), among persons having at least one visit in inpatient care.

Patient pathways from positive test to 30 days after

Overall, both men and women in all age groups largely followed the same patient pathways in the first two waves (Fig 6). The share seeking outpatient care decreased for those aged 1–19 (from 77% [95%CI = 76%-78%] to 63% [63%-64%]) and 20–67 (from 79% [78%-82%] to 72% [71%-72%]) from the first two waves to the third wave, respectively, while this remained stable from the first two waves to the third wave for those aged 68+ (57% [55%-59%]). In the fourth wave, the decrease in outpatient health care use continued for those aged 1–19 (51% [51-52]), 20–67 (63% [62%-63%]), and now also for those aged 68+ (50% [48%-52%]) (Fig 6).

Fig 6

Health care use.

Health care use during the first 30 days after a positive PCR-test for SARS-CoV-2 by age groups and gender, for four waves/periods of the pandemic: February 26th, 2020 –February 16th, 2021 (1st period), February 17th, 2021 –July 10th, 2021 (2nd period), July 11th, 2021 –December 27th, 2021 (3rd period), and December 28th, 2021 –January 14th, 2022 (4th period).

Along this decrease in outpatient care use, we also observed a decrease in the share of persons in need for inpatient care, especially from the 2nd to 4th wave for all groups of age and sex (Fig 6). For patients aged 68+ the share hospitalized declined from 25% [23%-27%] the first two ways, to 7% [5%-8%] in the fourth wave. The share that did not use any health care services (neither inpatient or outpatient) was relatively stable in wave 1 and 2 for all age groups and both sexes, but increased in wave 3, and continued to increase in wave 4 (Fig 6).

The shares that were still in need of outpatient or inpatient care on day 30 after a positive test decreased from the 1st to the 2nd wave and to a lesser extent from the 2nd to the 3rd wave, for all groups of age and sex. During the 1st wave, a smaller shared of those aged ≥68 years relied on self-care on days 21–30 after a positive test, compared to what is seen during the 3rd wave and 4th wave (S1 Fig).

And finally, the overall 30-day all-cause mortality decreased for all persons from the 1st to the 4th wave (S1 Table).

Discussion

In this descriptive study of all 493 520 persons testing positive for SARS-CoV-2 from February 26th, 2020 to January 14th, 2022 in Norway, we find that the share of patients with COVID-19 in need for health care services in the acute phase (30 days) has decreased throughout the pandemic, from 83% [95%CI = 83%-84%] to 59% [59%-59%] for outpatient care and from 4.9% [95%CI = 4.7%-5.0%] to 0.5% [0.4%-0.5%] for inpatient care. Accordingly, the share relying on self-care only increased, from 16% [15%-16%] in the first wave to 41% [41%-41%] in the fourth wave.

Comparison with previous studies

To our knowledge, this study is the first to describe COVID-19 related patterns in both inpatient and outpatient care use following positive test for SARS-CoV-2. The massive use of outpatient care services (including visits to the general practitioner and emergency wards) following COVID-19 has not previously been reported, i.e., no knowledge foundation exists for an effective comparison of findings across studies.

Far more is known for inpatient care use. Compared to previous studies we find, on average, lower hospitalization rates in Norway than what has been reported in other countries [20-22]. As an example, Denmark and Sweden observed hospitalization rates of 20% and 16%, respectively, in the first wave of the pandemic [21, 22], which contrasts with Norwegian hospitalization rates of only 5% (Table 1). The differences may be explained by differences in test criteria, but also differences in criteria for admission to the hospital. Further, the Danish and Swedish rates are based on a pandemic period that was shorter than studied in the current study [21, 22]. There were also differences in mortality between the Scandinavian countries. Whereas all-cause mortality among COVID-19 patients in Sweden was observed to be around 4–5% [22], we find a mortality rate of only 0.9% in Norway between February 2020 and February 2021 (Table 1). Again, the discrepancies might be explained by the difference in length of the study periods as well as differences in registration practices of deaths. However, in line with previous studies, we see higher mortality rates among men compared to women, particularly for those younger than 68 years [10]. Also, in line with previous reports, we find that a significant share of those who died, died within the first 10 days (S1 Fig) [23].

Another important observation to our inpatient care results, was the tendencies that the mean number of bed-days in hospital increased for some age- and sex-groups from the 1st to the 2nd wave (Fig 5). Although we did not aim to explore whether the severity of COVID-19 has changed, an important characteristic of the 2nd, 3rd, and 4th waves of transmission has been the rise of mutant viruses. Reports have been inconclusive as to whether mutant viruses (Alpha/Beta/Gamma vs. Wuhan strain) result in more severe disease requiring more hospital care [19, 24, 25]. Whilst the risk of hospitalization between the Alpha and Delta variant were found similar, the Omicron variant has had a reduced risk of hospitalization compared to the Delta variant [8, 9].

Interpretation and relevance

The high burden put on outpatient care services is important to report, given that a well-functioning outpatient care service is essential in reducing demands put on hospital services; it is essential to support rehabilitation of recovering patients; to improve palliative care; and sustain non-covid care [26]. As an example, S1 Fig shows that a larger share used inpatient care prior to outpatient care, than the other way around, suggesting that outpatient care to some extent has been used for recovery issues after hospitalization. As such, more knowledge of rehabilitation of COVID-19 patients (severely and mildly affected) in the community care services is needed.

Still, in our study, the peak use of outpatient care was centered to the -2 to 0 days around positive test, implying that a certain proportion of the large amount of outpatient care visits took place in relation to testing and the detection of COVID-19. Indeed, when we excluded visits that were related to testing, the total share visiting outpatient care during the 30-day period decreased, yet only slightly (from ~80% to ~70%). Thus, the somewhat different patterns in outpatient care use from the 1st and 2nd to the 3rd and 4th wave (Fig 2) may be explained by differences in testing criteria, i.e., the start of the 1st wave was the only wave with limited test availability and strict testing criteria (the elderly, persons at risk and health personnel). We believe that such differences in testing patterns are less likely to explain the decreasing demand put on inpatient care services throughout the pandemic (because frail and elderly persons were tested to an equal extent independent of pandemic wave).

A last interpretation of our findings is that healthcare workers in outpatient care may have a higher threshold for referral to inpatient care in 2022 than in 2020. This may be explained by increased knowledge of the disease and its outcomes (including lower mortality rates) as the pandemic progressed. The milder SARS-CoV-2 omicron variant and a higher vaccination coverage might also explain the observed shifts in health care use (Fig 1). However, we cannot exclude that more severe mutations in the future again place an increasing demand on the health care services. Also, vaccines might have lower effects against new variants. Thus, if we again see SARS-CoV-2 variants that cause the same disease severity as the initial variants, our findings of an only halved outpatient care use but ten times lower inpatient care use from February 2020 to February 2022 suggest that an upscaling of the outpatient care services might be particularly important in the future.

Strength and limitations

An important strength of our study is that we could include everyone with a positive test throughout four major waves or periods of the pandemic. In this way, we could provide a comprehensive picture of all health care use following a positive PCR-test for these different waves, i.e., not restricted to inpatient care as in previous studies [2, 3]. Moreover, we could provide details in outpatient and inpatient care for different age and sex groups.

Several limitations should be mentioned. First, we do not know the causes or severity of complaints behind the care use following a positive test for SARS-CoV-2. Although we only included care visits with diagnostic codes of COVID-19, we did not separate the complaints affecting e.g., the respiratory or digestive system. Also, we had no comparison group, simply because we did not aim for any causal inference and because comparable data are not available for a similar epidemic or pandemic setting with other infectious diseases. However, in recent studies of post-acute COVID-19, we demonstrate a likely causal effect of being infected with SARS-CoV-2 on the post-acute health care use [27]. Here, we also exclusively included visits that were specific to COVID-19, i.e., we did not study all-cause visits. Second, our study was of an explorative and descriptive character. Thus, we looked for patterns and trends in a large amount of data using mainly graphs in a self-developed structure, such as the division of age into children and adolescents, adults in working age population and the elderly, and by sex. We did not apply any data-driven analyses in our exploration of pandemic trends in health care use, thus we might have missed important details. To combat some of these issues, we chose to present a large amount of raw data visualized as alluvial diagrams in the S1 Fig. Third, we may have underestimated the care use among persons aged 68 years or more. Very frail persons live in care homes and receive institutionalized care that may not be registered in our data sources. And finally, and as mentioned above, we cannot exclude that some of our observations of changing (or stable) trends are due to differences in test criteria or -patterns as the pandemic progressed. Such patterns may differ across our groups of age and sex. However, if this is the case, the testing is obviously a part of health care use in relation to COVID-19, or else we would not have observed these visits. Thus, because testing for SARS-CoV-2 has been a part of the outpatient and inpatient care services from the beginning of the pandemic, including care visits in relation to the detection of COVID-19 is still important in the public health question of whether the health services should be upscaled or downscaled in future similar situations.

In conclusion, we demonstrate a decreasing impact of COVID-19 on all COVID-19 related health care services from the 1st to the 4th wave of the pandemic. The use of COVID-19 related outpatient care services was reduced with 29%, whereas the use of COVID-19 related inpatient care services was reduced with 80% in January-February 2022 compared to the first year of the pandemic. These findings are important to report considering future mutations and waves of COVID-19, i.e., there may be a lower need for upscaling inpatient care services and a large need for upscaling outpatient care services.

Alluvial diagram showing patient-flows in ten day-intervals by age group and period.

Note: This figure visualizes the most common pathway-flows for each period and each age-group. SC = Self-care; OC = Outpatient care; IC = Inpatient care; and Dead = Dead within 30 days. Colors show which strata each person belonged to in the last 10 days. Categories are mutually exclusive and ordered by severeness (Death > IC > OC > SC). Crude numbers are removed in order to avoid confusion. The sizes of each stratum should therefore be interpreted as shares.

(PDF)

Click here for additional data file.

Share visiting outpatient care withing 30 days of testing positive by age and sex.

Note: Day 0 includes 0–2 days before testing positive.

(PDF)

Click here for additional data file.

Day by day cumulative share visiting outpatient care (GP or emergency ward) from one day after testing positive (day 1) to 30th day after positive test, i.e., excluding outpatient care visits that were related to the testing and detection of SARS-CoV-2.

(PDF)

Click here for additional data file.

Share visiting inpatient care withing 30 days of testing positive by age and sex.

Note: Day 0 includes 0–2 days before testing positive.

(PDF)

Click here for additional data file.

Day by day cumulative share visiting inpatient care from one day after testing positive (day 1) to 30th day after positive test, i.e., excluding inpatient care visits that were related to the testing and detection of SARS-CoV-2.

(PDF)

Click here for additional data file.

Outcomes of persons testing positive for SARS-CoV-2 within 30 days, in each of four pandemic waves in Norway, 2020–2022.

Note: Due to privacy reasons we cannot report exact numbers when numbers are between 0 and 5. Therefore we have censured the exact numbers for deaths in the 1st and 3rd wave for the various age groups. The table still includes the percentages and 95% confidence intervals.

(PDF)

Click here for additional data file.

10 Jan 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Mabel Aoun, MD, MPH

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. Your ethics statement should only appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please delete it from any other section.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Thanks to the authors. I reviewed the article. It is valuable research in the pandemic era. The title, introduction, methodology, and conclusion are appropriate. There is no plagiarism in it. The writing is appropriate and completely understandable. I have minor recommendation as followings:

1- How did the authors delete the effect of confounding factors from the study outcome?

2- Discussion should be improved and revised based on the study results and comparison with previous ones i.e. Bagi HM, Soleimanpour M, Abdollahi F, Soleimanpour H (2021) Evaluation of clinical outcomes of patients with mild symptoms of coronavirus disease 2019 (COVID-19) discharged from the emergency department. PLoS ONE 16(10): e0258697. https://doi.org/10.1371/journal.pone.0258697

3- What is the new finding of this study compared to previous ones this field?

4- Please mention the weak and strong points of your study

5- The list of abbreviations at the end of the manuscript is absence; please make sure all your abbreviations are listed.

6- I recommend providing a table on the characteristics of your sample, including all variables you use in your analysis later on.

7- Please describe the study design in more detail.

8- provide more information on how you analyzed your data and possibly provide some references for e.g. level of statistical significance.

Reviewer #2: Introduction (pages 3-4): Please state study design (national cohort study?) and any prespecified hypotheses. More generally, please review and consider using STROBE checklist for observational studies (e.g. https://www.strobe-statement.org/checklists/).

Methods (page 4): An important feature of this study is its use of Norwegian national registries. The authors give one reference to a webpage that describes the BEREDT-C19 registry. I’m unable to find any validation data on that website. If such validation data exists, the authors should cite it directly. If validation data of the BEREDT-C19 registry does not exist but validation of the underlying registries from which BEREDT-C19 draws does exist, I would cite that data instead. The validity of these registries may be well known to Norwegian epidemiologists, but should be established for international readers.

Methods (page 5-6): It becomes apparent here that the authors are including treatment received in emergency wards as “primary care,” as contrasted with “specialist care” meaning hospital admission. I personally wouldn’t usually think of emergency department visits as part of primary care, and I don’t think most US physicians would either. Conversely, specialist care doesn’t clearly connote hospitalization to me (I would think of e.g. an outpatient cardiology clinic visit as “specialist care”). If I’m understanding the authors’ intentions correctly, I wonder if the manuscript might be clearer for international readers if it consistently referred to “outpatient care” (including clinic and emergency department visits) vs. “inpatient care” (hospital admission) rather than to “primary care” vs. “specialist care?”

Methods (page 6): the authors describe the three waves they are examining and mention the start of mass vaccination during the third wave. It would be helpful to have numbers and citations for the rates of vaccination at the beginning and end of this wave, since this would be expected to significantly impact rates of hospitalizations and I would think this data would be relatively easy to obtain.

Results (pages 7-11): The authors don’t have a comparison group but make many comparisons between waves, e.g. “we observed a significant shift in the total use of specialist care throughout the pandemic, from 14% being hospitalized at least once during the 1st wave, compared to 4% during the 2nd and 3rd waves (Fig 4).” (page 10) Consider measures of statistical significance or confidence intervals for differences.

Discussion (page 12): The authors state that “we were unable to find a similar study for an effective comparison of our findings.” I agree that I also am not able to find a similar study of the proportion of COVID patients receiving outpatient care. However, as the authors note, there are many studies of hospitalization rates among patients with COVID (e.g. Menachemi N, Dixon BE, Wools-Kaloustian KK, Yiannoutsos CT, Halverson PK. How Many SARS-CoV-2-Infected People Require Hospitalization? Using Random Sample Testing to Better Inform Preparedness Efforts. J Public Health Manag Pract. 2021 May-Jun 01;27(3):246-250. doi: 10.1097/PHH.0000000000001331. PMID: 33729203.). The authors should discuss how their findings regarding hospitalization rates compare to those found in other studies (the hospitalization rates the authors found appear to me higher than those found in some other studies).

Discussion (page 13): The authors note that the proportion of patients hospitalized and the duration of hospitalization increased from wave 2 to wave 3. Given that the authors state that vaccines became available during wave 3, these findings are surprising. Again, it would be helpful to know what proportion of the population was vaccinated at the beginning and end of the 3rd wave. Some comment might be offered in the discussion.

Finally, the authors may consider including data from the current omicron wave. I don’t think it’s absolutely necessary, and the publication need not and should not be delayed until the pandemic fully passes, but the study would be even more informative with data from this fourth wave included.

Reviewer #3: This study highlights a major topic of interest during the Covid 19 pandemic.

The healthcare burden imposed by the pandemic is one of the main factors to be studied in order to improve preparedness plans for the coming waves and to enhance knowledge about dealing with future pandemics. This study deals with exhaustive data from a national registry and is theoretically well positioned to give valuable information about the research question risen by the authors.

Understanding patients flow during the pandemic is another important facet of managing healthcare resources knowing that the study reports data from Norway, a country known for a very distinguished and organized health system.

The authors decided to study sex and age specific use of healthcare system, I would appreciate adding to the introduction the rationale behind this choice as they were only mentioned in the research question without any evidence to support this choice.

The 30 days’ time frame is also a choice made by the authors and showing evidence supporting this time limit -mean time to recovery, incidence of late onset complications…- would also be of value.

The authors developed in a descent way the importance of their research question. Developing the potential applications of such data would add value to the manuscript.

The outcomes measures are not well described in the methods section. Primary care use for example is a broad topic and the indicators used for measurement developed later may figure in a paragraph with a detailed description of the calculation methods used for every variable used to assess this outcome.

A review of the editing and English proofing would also be helpful to make the reading even more enjoyable.

Lines 59-72: the description of the registry is fair enough and gives a good insight of its contents and objectives. A clarification about its exhaustiveness and any possible missed data is important to know. The ethical committee

Line 86 : are there any other reasons for confusion factors? Do we have data on coding accuracy?

Line 88: Did the authors check for a more general coding Like R99?

Lines 111-119: definition of cumulative and peak use of care would fit better in the outcomes section. I would like the authors to clarify the discrepancy between primary care and specialist care representation where they included the first visit only for primary care while they calculated the hospital bed days.

Lines 126-135: I suggest the overall results being presented first before going into the different subgroups analysis to make the reading easier and to be in accordance with the primary objective of the study and the analysis plan described in the statistical analysis section

Lines 157-159: I couldn’t see on this figure that the share in need of specialist care prior to primary care was larger than its inverse. It may need review and clarification.

Death data was not presented in the results section

the term "Share" and "Fraction" are used mutually in different locations creating confusion. i suggest to authors using one term throughout the article.

The discussion is very well structured. It states the main results with a global view on the objectives of the study and discuss the possible reasons of the observed trends while suggesting hypotheses to test in future research. The limitations are well discussed making the conclusions drawn from this exploratory analysis reasonable and sound.

Figures are difficult to read because of the low resolution. Y axis unit is not shown for fig 1,2 and 4

The design of figure 1 makes the comparison of the different sex and age groups not easy.

The S1 Fig shows the absolute numbers on the y axis which makes any visual comparison erroneous.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Benjamin Tolchin

Reviewer #3: Yes: Marouan Zoghbi

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

2 Mar 2022

(For more intuitive formatting see the uploaded file with response to reviewers.)

We would like to thank the expert reviewers for valuable input, which has helped to improve the quality of our manuscript. Please find below a point-to-point response to your comments and a list of the changes we made in the revised manuscript.

Comments of Academic Editor

1-Describe the study design, make sure your paper follows the STROBE checklist for observational studies

We agree. We incorporate a sentence where we describe the study design as a national cohort study using registry data, p. 3, lines 50-52: “Thus, in this paper we aim to explore the age- and sex-specific acute burden of COVID-19 on the health care services in four waves of the pandemic in Norway through a national cohort study using registry data.”We have also attached a STROBE checklist.

2-Please address the comments of reviewer #2 regarding the BEREDT-C19 registry and its validation data. What are the flaws of the registry, missing data?

We agree that issues such as flaws of the registry including missing data should have been better described. We have now described the flaws of the registry (such as non-quality checked data) in the methods section, p. 4-5 and lines 77-80: “Overall, data from Norwegian health registers have been demonstrated to be of high quality with high validity and reliability, and together they can provide a complete picture of patterns of healthcare use [11-13]. Medical recording to the National registries is mandated by law in Norway, ensuring no missing data in our study.”

3-Clearly define the criteria for primary care and specialist care, as highlighted by reviewers #2 and #3.

Yes, the definitions of primary care and specialist visits could have been clearer. To make this clearer to the reader we have changed from “primary care” to “outpatient care”, and from “specialist care” to “inpatient care”, as suggested by the reviewers.

4-Please explain the rationale behind using age- and sex-specific share of patients to assess healthcare system use and the rationale behind the 30-day acute phase as highlighted by reviewer #3

We agree this could have been better described in the background section. Age and sex differences were already well-known, e.g. that children are less severely affected than elderly in terms of hospital admissions. Here, we aimed to shed light on admissions in combination with primary care visits for the different age and sex groups. We have added the following to the introduction, p. 3-4, lines 52-56): “Because of previously reported differences in vaccination status throughout the pandemic [7], differences in disease severity by SARS-CoV-2 variant [8, 9] and strata of age and sex [10], we hypothesized that we would also see differences in the 30-day pattern of healthcare use for men and women, girls and boys, the working age population and the elderly in the different waves of the pandemic.” And the following sentence to the methods section, p. 5, lines 87-90: “We divided our population into mutually exclusive age and sex groups, i.e. girls and boys, men and women by the following age categories: 1-19 (children and adolescents), 20-67 (working age population) and 68 years or older (elderly), as COVID-19 has hit differently among different age groups and sexes [10].” And the following to the methods section, p. 6, lines: 101—104: “We chose a 30-days-timeframe because a death after COVID-19 was classified as covid-related if it occurred within 30 days after testing positive in official statistics [15], and because it coincides with what is commonly referred to as the acute phase of SARS-CoV-2 [16].”

5-Clarify figure S1

We agree. We have now removed the numbers on the Y-axis in Figure S-1 to remove any confusion, and we have added more information in the figure notation.

6-Please use one term either share or fraction

We agree. We have revised our manuscript throughout to use the term “share” instead of “fraction”.

Comments of reviewer 1

Thanks to the authors. I reviewed the article. It is valuable research in the pandemic era. The title, introduction, methodology, and conclusion are appropriate. There is no plagiarism in it. The writing is appropriate and completely understandable. I have minor recommendation as followings:

Thank you for your encouraging comments.

1. How did the authors delete the effect of confounding factors from the study outcome?

In our study, we aimed to simply describe the health services use in primary and specialist care (now called outpatient and inpatient care) following a positive test for SARS-CoV-2. We did not aim to test a causal hypothesis, and thus, there was no need to control for any potentially confounding factors. However, we agree that factors impacting on health care use, such as SARS-CoV-2 variant and vaccination could have been described in our paper, although there was no need to adjust for them in the statistical analyses. Besides more clearly describing that we had a descriptive aim and providing a clearer background for a broad hypothesis of what we expected to find in the descriptive data, we now include more data on SARS-CoV-2 variants and vaccination grade, please see Figure 1. We also include more of this information in our discussion section.

2- Discussion should be improved and revised based on the study results and comparison with previous ones i.e. Bagi HM, Soleimanpour M, Abdollahi F, Soleimanpour H (2021) Evaluation of clinical outcomes of patients with mild symptoms of coronavirus disease 2019 (COVID-19) discharged from the emergency department. PLoS ONE 16(10): e0258697. https://doi.org/10.1371/journal.pone.0258697

Thank you for this valuable input. We have included a more thorough discussion and included the suggested references. We have also improved the structure of our entire discussion section, with headings “Comparison to previous studies”, “Interpretation and relevance” and “Strengths and limitations” (edits too long to be pasted here).

3- What is the new finding of this study compared to previous ones this field?

We agree that we could have better described the independent contribution relative to previous studies. Amongst others, we have made the following revisions to our discussion section, p. 13, lines 252-264:

“To our knowledge, this study is the first to describe COVID-19 related patterns in both inpatient and outpatient care use following positive test for SARS-CoV-2. The massive use of outpatient care services (including visits to the general practitioner and emergency wards) following COVID-19 has not previously been reported, i.e. no knowledge foundation exists for an effective comparison of findings across studies. Far more is known for inpatient care use. Compared to previous studies we find, on average, lower hospitalization rates in Norway than what has been reported in other countries [20-22]. As an example, Denmark and Sweden observed hospitalization rates of 20% and 16%, respectively, in the first wave of the pandemic [21, 22], which contrasts with Norwegian hospitalization rates of only 5% (Table 1). The differences may be explained by differences in test criteria, but also differences in criteria for admission to the hospital. Further, the Danish and Swedish rates are based on a pandemic period that was shorter than studied in the current study [21, 22]. There were also differences in mortality between the Scandinavian countries.”

4- Please mention the weak and strong points of your study

We agree that the weak and strong points of our study could have been better structured. Please see our section of the discussion section entitled “Strengths and limitation”, p. 15-17, lines 308-342.

5- The list of abbreviations at the end of the manuscript is absence; please make sure all your abbreviations are listed.

We see that this could be useful, however we could not find that this is standard practice in this journal. We will add a list of abbreviations if required by the journal.

6- I recommend providing a table on the characteristics of your sample, including all variables you use in your analysis later on.

We agree. We have enlarged Table 1 with more information, and we now also include a new table (S1-Table) in the appendix showing the numbers and percentages of all health care use for each sub-groups and waves.

7- Please describe the study design in more detail.

We agree. We have added the following to the introduction section, p. 3, lines 50-52: “Thus, in this paper we aimed to explore the age- and sex-specific acute burden of COVID-19 on the health care services in four waves of the pandemic in Norway through a national descriptive cohort study design using registry data.”

8- provide more information on how you analyzed your data and possibly provide some references for e.g. level of statistical significance.

We agree. We have added the following to the statistical analyses section, p. 8, lines 138-140: “For each age-/sex-group and for each of their pathways in each of their pandemic wave, we estimated the proportions having the different pathways and calculated 95% confidence intervals.” Accordingly, we have updated our Figure 6 with the confidence intervals.

Comments of reviewer 2

Introduction (pages 3-4): Please state study design (national cohort study?) and any prespecified hypotheses. More generally, please review and consider using STROBE checklist for observational studies (e.g. https://www.strobe-statement.org/checklists/).

Thank you. We have added the following to the introduction section, p. 3, lines 50-52: “Thus, in this paper we aimed to explore the age- and sex-specific acute burden of COVID-19 on the health care services in four waves of the pandemic in Norway through a national descriptive cohort study design using registry data.” Our study design was of a descriptive and explorative nature, and we had no prespecified hypothesis in terms of direction of results. However, as rationale for our choice of groupings and categorizations, we have added the following hypotheses, p. 3-4, lines 52-56: “Because of previously reported differences in vaccination status throughout the pandemic [6], differences in disease severity by SARS-CoV-2 variant [7, 8] and strata of age and sex [9], we hypothesized that we would also see differences in the 30-day pattern of healthcare use for men and women, girls and boys, the working age population and the elderly in the different waves of the pandemic.” We have also attached a STROBE checklist.

Methods (page 4): An important feature of this study is its use of Norwegian national registries. The authors give one reference to a webpage that describes the BEREDT-C19 registry. I’m unable to find any validation data on that website. If such validation data exists, the authors should cite it directly. If validation data of the BEREDT-C19 registry does not exist but validation of the underlying registries from which BEREDT-C19 draws does exist, I would cite that data instead. The validity of these registries may be well known to Norwegian epidemiologists, but should be established for international readers.

Thanks for pointing out. We agree that this would greatly benefit the paper. We now include a sentence on the quality of the underlying data of the register, together with a reference. See methods section, p. 4-5 and lines 77-80: “Overall, data from Norwegian health registers have been demonstrated to be of high quality with high validity and reliability, and together they can provide a complete picture of patterns of healthcare use [11-13]. Medical recording to the National registries is mandated by law in Norway, ensuring no missing data in our study.”

Methods (page 5-6): It becomes apparent here that the authors are including treatment received in emergency wards as “primary care,” as contrasted with “specialist care” meaning hospital admission. I personally wouldn’t usually think of emergency department visits as part of primary care, and I don’t think most US physicians would either. Conversely, specialist care doesn’t clearly connote hospitalization to me (I would think of e.g. an outpatient cardiology clinic visit as “specialist care”). If I’m understanding the authors’ intentions correctly, I wonder if the manuscript might be clearer for international readers if it consistently referred to “outpatient care” (including clinic and emergency department visits) vs. “inpatient care” (hospital admission) rather than to “primary care” vs. “specialist care?”

Thank you. We agree. We now use the term “outpatient care” instead of “primary care”, and “inpatient care” instead of “specialist care”.

Methods (page 6): the authors describe the three waves they are examining and mention the start of mass vaccination during the third wave. It would be helpful to have numbers and citations for the rates of vaccination at the beginning and end of this wave, since this would be expected to significantly impact rates of hospitalizations and I would think this data would be relatively easy to obtain.

We agree. We now incorporate a graph showing the share of persons vaccinated with at least one dose (Fig 1). Moreover, we include this in the Discussion, see p. 15, lines 300-306: “The milder SARS-CoV-2 omicron variant and a higher vaccination coverage might also explain the observed shifts in health care use (Fig 1). However, we cannot exclude that more severe mutations in the future again place an increasing demand on the healthcare services. Also, vaccines might have lower effects against new variants. Thus, if we again see SARS-CoV-2 variants that cause the same disease severity as the initial variants, our findings of an only halved outpatient care use but ten times lower inpatient care use from March 2020 to February 2022 suggest that an upscaling of the outpatient care services might be particularly important in the future.”

Results (pages 7-11): The authors don’t have a comparison group but make many comparisons between waves, e.g. “we observed a significant shift in the total use of specialist care throughout the pandemic, from 14% being hospitalized at least once during the 1st wave, compared to 4% during the 2nd and 3rd waves (Fig 4).” (page 10) Consider measures of statistical significance or confidence intervals for differences

We agree. We now include a table (S-Table 1) showing all rates with corresponding 95% confidence interval, to show whether the differences in health care use across the various wave are significant or not. We have also updated Fig 6 showing 95% confidence intervals.

Discussion (page 12): The authors state that “we were unable to find a similar study for an effective comparison of our findings.” I agree that I also am not able to find a similar study of the proportion of COVID patients receiving outpatient care. However, as the authors note, there are many studies of hospitalization rates among patients with COVID (e.g. Menachemi N, Dixon BE, Wools-Kaloustian KK, Yiannoutsos CT, Halverson PK. How Many SARS-CoV-2-Infected People Require Hospitalization? Using Random Sample Testing to Better Inform Preparedness Efforts. J Public Health Manag Pract. 2021 May-Jun 01;27(3):246-250. doi: 10.1097/PHH.0000000000001331. PMID: 33729203.). The authors should discuss how their findings regarding hospitalization rates compare to those found in other studies (the hospitalization rates the authors found appear to me higher than those found in some other studies).

Thank you for suggesting relevant literature. We now include a more thorough discussion on how our findings relate to previous literature on COVID-19 patients seeking inpatient health care. We include the provided references. Please see our response and action to the comments above and our new section included in the manuscript, entitled “Comparison with previous studies”, p. 13-14 (too lengthy to be pasted here).

Discussion (page 13): The authors note that the proportion of patients hospitalized and the duration of hospitalization increased from wave 2 to wave 3. Given that the authors state that vaccines became available during wave 3, these findings are surprising. Again, it would be helpful to know what proportion of the population was vaccinated at the beginning and end of the 3rd wave. Some comment might be offered in the discussion.

We agree that showing the proportion of vaccinated would improve the paper. We now include a graph (Fig 1) showing the share of the population vaccinated with at least one dose throughout the various waves. We also include the following section in the discussion section, p. 14, lines 271-278: "Another important observation to our inpatient care results, was the tendencies that the mean number of bed-days in hospital increased for some age- and sex-groups from the 1st to the 2nd wave (Fig 5). Although we did not aim to explore whether the severity of COVID-19 has changed, an important characteristic of the 2nd, 3rd, and 4th waves of transmission has been the rise of mutant viruses. Reports have been inconclusive as to whether mutant viruses (Alpha/Beta/Gamma vs. Wuhan strain) result in more severe disease requiring more hospital care [19, 24, 25]. Whilst the risk of hospitalization between the Alpha and Delta variant were found similar, the Omicron variant has had a reduced risk of hospitalization compared to the Delta variant [8, 9].”

Finally, the authors may consider including data from the current omicron wave. I don’t think it’s absolutely necessary, and the publication need not and should not be delayed until the pandemic fully passes, but the study would be even more informative with data from this fourth wave included.

We agree that this would greatly improve the paper. We have now extended the study period until 14th of February 2022 (last positive test on 14th of January 2022), to capture data on the latest wave.

Comments of reviewer 3

This study highlights a major topic of interest during the Covid 19 pandemic. The healthcare burden imposed by the pandemic is one of the main factors to be studied in order to improve preparedness plans for the coming waves and to enhance knowledge about dealing with future pandemics. This study deals with exhaustive data from a national registry and is theoretically well positioned to give valuable information about the research question risen by the authors.

Understanding patients flow during the pandemic is another important facet of managing healthcare resources knowing that the study reports data from Norway, a country known for a very distinguished and organized health system.

The authors decided to study sex and age specific use of healthcare system, I would appreciate adding to the introduction the rationale behind this choice as they were only mentioned in the research question without any evidence to support this choice.

Thank you for your encouraging feedback. We decided to study sex and age specific use as healthcare use have differed between the two sexes and for various age groups. A 30-day frame have often been considered as the acute phase. Anything beyond that is often considered as “long covid”, which is not the topic of this article. We agree that the rationales for our choices could have been better described. We have added the following to the introduction, p. 34, lines 52-56): “Because of previously reported differences in vaccination status throughout the pandemic [6], differences in disease severity by SARS-CoV-2 variant [7, 8] and strata of age and sex [9], we hypothesized that we would also see differences in the 30-day pattern of healthcare use for men and women, girls and boys, the working age population and the elderly in the different waves of the pandemic.” And the following sentence to the methods section, p. 5, lines 87-90: “We divided our population into mutually exclusive age and sex groups, i.e. girls and boys, men and women by the following age categories: 1-19 (children and adolescents), 20-67 (working age population) and 68 years or older (elderly), as COVID-19 has hit differently among different age groups and sexes [9].”

The 30 days’ time frame is also a choice made by the authors and showing evidence supporting this time limit -mean time to recovery, incidence of late onset complications…- would also be of value.

We agree. We have added the following to the introduction, p. 3-4, lines 52-56): “Because of previously reported differences in vaccination status throughout the pandemic [6], differences in disease severity by SARS-CoV-2 variant [7, 8] and strata of age and sex [9], we hypothesized that we would also see differences in the 30-day pattern of healthcare use for men and women, girls and boys, the working age population and the elderly in the different waves of the pandemic.” And the following to the methods section, p. 6, lines: 101—104:“We chose a 30-days-timeframe because a death after COVID-19 was classified as covid-related if it occurred within 30 days after testing positive in official statistics [15], and because it coincides with what is commonly referred to as the acute phase of SARS-CoV-2 [16]. Thus, we regarded people who were still alive after 30 days as recovered.”

The authors developed in a descent way the importance of their research question. Developing the potential applications of such data would add value to the manuscript

We agree. We now incorporate a sentence on what the data may be used to in other studies, on p. 3, lines 43-48: “A timely and correct up- and downscaling of health services (and lockdown measures) depend on our understanding of the pathways patients take through the health system, including the peaks and total demand of health care services following an individual’s positive test. Studying the impact on both inpatient and outpatient care in its early waves, can provide valuable insight into health service needs in later stages, and contribute to the knowledge base that can increase our resilience against future pandemics.” These and other revisions contributed to a significantly increased word count of our introduction section, i.e. we have also deleted the two first sentences describing lockdown (line 36-40) as we think it was less relevant for our research question.

The outcomes measures are not well described in the methods section. Primary care use for example is a broad topic and the indicators used for measurement developed later may figure in a paragraph with a detailed description of the calculation methods used for every variable used to assess this outcome.

We agree. We now use the term ‘outpatient’ and ‘inpatient’ instead of ‘primary’ and ‘specialist’ care, respectively. We have also restructured our outcomes section, placing the specific outcomes first, and how they were handled in the analyses thereafter.

A review of the editing and English proofing would also be helpful to make the reading even more enjoyable.

We agree. We have edited the English language.

Lines 59-72: the description of the registry is fair enough and gives a good insight of its contents and objectives. A clarification about its exhaustiveness and any possible missed data is important to know. The ethical committee

We agree. We have added the following to p. 4-5, lines 77-80: “Overall, data from Norwegian health registers have been demonstrated to be of high quality with high validity and reliability, and together they can provide a complete picture of patterns of healthcare use [11-13]. Medical recording to the National registries is mandated by law in Norway, ensuring no missing data in our study.”

Line 86 : are there any other reasons for confusion factors? Do we have data on coding accuracy?

In our study, we aimed to simply describe the health services use in primary and specialist care (now called outpatient and inpatient care) following a positive test for SARS-CoV-2. We did not aim to test a causal hypothesis, and thus, there was no need to control for any potentially confounding/confusion factors. However, we agree that factors impacting on health care use, such as SARS-CoV-2 variant and vaccination could have been described in our paper, although there was no need to adjust for them in the statistical analyses. We agree that we could have included more information regarding coding accuracy. With regards to confusion factors, we have provided a better description of our study design as well as a rationale for our research aim and hypothesis, please see the introduction section, p. 3-4, lines 52-56. Further, we provide data on factors impacting on health care use, such as vaccination and SARS-CoV-2 variant in Fig 1. Regarding coding accuracy, please see response above.

Line 88: Did the authors check for a more general coding Like R99?

The more general code R99 was considered, however as it represents any “respiratory tract infection not classified elsewhere”, i.e. not restricted to covid-related healthcare use, we decided to avoid it in our study.

Lines 111-119: definition of cumulative and peak use of care would fit better in the outcomes section. I would like the authors to clarify the discrepancy between primary care and specialist care representation where they included the first visit only for primary care while they calculated the hospital bed days.

We agree that peak and total use of healthcare could have been described in the Outcomes section. However, we would like to keep the details of our estimation strategies for the peak and total healthcare use in the statistical analyses section. We have added the following to the section on outcomes, page 6, lines 105-107: “When combined, and sorted chronologically on dates of occurrence, these data provided a comprehensive picture of the peak and total use of outpatient and inpatient care, as well as COVID-19-related health care pathways in the acute phase.” Moreover, we now include all outpatient care visits when estimating peak and total use, not only the first visit. This also makes it more comparable to the way we measure inpatient care, as we include all bed-days and not just the admission date.

Lines 126-135: I suggest the overall results being presented first before going into the different subgroups analysis to make the reading easier and to be in accordance with the primary objective of the study and the analysis plan described in the statistical analysis section

Thank you for the suggestion. We agree. We now present the overall results before presenting the different subgroup analyses (patient pathways). We have revised the chronology of the statistical analyses section accordingly.

Lines 157-159: I couldn’t see on this figure that the share in need of specialist care prior to primary care was larger than its inverse. It may need review and clarification. We believe that it is even clearer now. In S1 Fig, when looking at the top right figure, the share going from inpatient care (IC) to outpatient care (OC) is larger than the share going from outpatient care (OC) to inpatient care (IC).

Death data was not presented in the results section.

Death data were presented in the first paragraph in the Results section. However, we agree that these data should be described more thoroughly (lines 149-152). We now present even more death data by the inclusion of all-cause mortality in Table 1 and more in-depth in S1 Table. We also describe mortality more thoroughly in the Discussion, p. 3, lines 264-270: “Whereas all-cause mortality among COVID-19 patients in Sweden was observed to be around 4-5% [22], we find a mortality rate of only 0.9% in Norway between February 2020 and February 2021 (Table 1). Again, the discrepancies might be explained by the difference in length of the study periods as well as differences in registration practices of deaths. However, in line with previous studies, we see higher mortality rates among men compared to women, particularly for those younger than 68 years [10]. Also, in line with previous reports, we find that a significant share of those who died, died within the first 10 days (S1 Fig) [23].”

the term "Share" and "Fraction" are used mutually in different locations creating confusion. i suggest to authors using one term throughout the article.

We agree. We now use the term “share” throughout the paper.

The discussion is very well structured. It states the main results with a global view on the objectives of the study and discuss the possible reasons of the observed trends while suggesting hypotheses to test in future research.

Thank you. In accordance with comments from Reviewer 1 and 2, we have improved our discussion section even further, by the inclusion of subheadings, such as “Comparison to previous studies”, “Interpretation and relevance”, and “Strengths and limitations”.

The limitations are well discussed making the conclusions drawn from this exploratory analysis reasonable and sound.

Thank you.

Figures are difficult to read because of the low resolution. Y axis unit is not shown for fig 1,2 and 4

Thanks for letting us now. Figures are now in higher quality.

The design of figure 1 makes the comparison of the different sex and age groups not easy

We agree that the figure layout could have been better. We have restructured the figure (now renamed Fig 6) in such a way that the periodical shifts in healthcare use are now visually evident. We have also attached the 95% confidence intervals.

The S1 Fig shows the absolute numbers on the y axis which makes any visual comparison erroneous.

We agree. We have now removed the numbers from the y-axis, so that it is easier to understand that comparison is of percentages and not absolute numbers, and we have also added more information to the figure notation.

Submitted filename: Point-by-point_plos_one.docx

Click here for additional data file.

4 Mar 2022

7 Mar 2022

Please remove the figure 1 from the introduction and integrate it within a paragraph entitled "study setting" in the methods, where you can describe in a sentence or two the SARS-CoV-2 context in Norway.

We agree. We removed Fig 1 from the introduction and placed it within a subsection entitled “Study setting” within the Methods section. Instead of adding one to two sentences on the SARS-CoV-2 context in Norway, we moved the paragraph describing the different waves from the subsection “Statistical analyses” to this new section, as we felt it was more logical to place the paragraph in the new subsection.

We also noticed that we had made an error on line 92. We do not only include hospitalizations with ICD-10 code U071 (confirmed COVID-19), but we also include U072 (suspected COVID-19). This is because these patients have already tested positive for SARS-CoV-2, and in the beginning of the pandemic, hospitals were not coherent in in the use of U071 or U072. This is now corrected in the uploaded manuscript.

Submitted filename: Point-by-point_plos_one_new.docx

Click here for additional data file.

9 Mar 2022

Pandemic trends in health care use: From the hospital bed to self-care with COVID-19

PONE-D-21-37160R2

Dear Dr. Methi,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Mabel Aoun, MD, MPH

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

14 Mar 2022

PONE-D-21-37160R2

Pandemic trends in health care use: From the hospital bed to self-care with COVID-19

Dear Dr. Methi:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Mabel Aoun

Academic Editor

PLOS ONE