Impacts of COVID-19 on long-term health and health care use

Design & data sources

Applying an observational pre-post design with comparison group, we utilized population-wide longitudinal registry data from Norway to estimate impacts of COVID-19 on health care utilization. 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 [11]. BeredtC19 compiles daily updated individual-level data from several registers, including 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 primary 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 COVID-19 in Norway with date of testing and test result, reported from all laboratories in Norway to MSIS 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. Institutional board review was conducted, and 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 resident of Norway on January 1^st 2020 - as well as everyone born in 2020 - who had been tested for the SARS-CoV-2 by a PCR-test from March 1^st to November 1^st 2020 (non-residents like tourist etc excluded). With outcome data from January 1^st 2020 through January 31^th 2021, we followed the patients for at least two months before and at least three months after the test date.

COVID-19

In addition to a comparison group of patients with only negative PCR-tests for SARS-CoV-2 (no COVID-19), we studied patients in two mutually exclusive groups:

1. Mild COVID-19, comprising all individuals with a positive PCR test who were not hospitalized or dead during the 14 days after test date.

2. Severe COVID-19, comprising all individuals with a positive PCR test who were hospitalized from 2 days before to 14 days after test date.

Both in the very few cases with more than one positive test (mild and severe COVID-19) and when there were more negative tests (no COVID-19), we used the first available test date for the patient.

Outcomes

We studied all-cause and cause-specific utilization of primary and specialist care. For the different health care levels, our all-cause outcomes were three categorical variables set to one if the person had 1) visited primary care (i.e. general practitioners or emergency wards) at least once during a week, 2) received hospital-based outpatient specialist care at least once during a week, and 3) received hospital-based inpatient specialist care at least once during a week. Each of the three outcome variables were defined analogously for health care utilization by specific causes or complaints, i.e. by the chapters in the International Classification of Primary Care (ICPC-2) and the International Classification of Diseases (ICD-10), as described in Table 1. Thus, for cause-specific outcomes, it was possible that a patient had utilized health care services for diagnoses covered by more than one chapter in the same week.

Statistical analyses

Starting with all-cause health care utilization at any level, at the primary care level and at the outpatient and inpatient care level, we studied the weekly change in health care utilization from 3 months prior to test week, to 6 months after test week for the persons with no COVID-19, with mild COVID-19 and with severe COVID-19. We calculated the rate of health services utilization in each of the three groups (no, mild and severe COVID-19) per calendar week, and presented the results as the mean of the weekly rates in periods of one and three months.

To study the impacts of COVID-19 on health care utilization, we estimated the change in the health care outcome in question from before to after the positive test for those with COVID-19, and compared it to the change in the health care outcome in question from before to after the negative test for those with no COVID-19. This difference-in-differences estimate [12] captures the change in the health care outcome from before to after COVID-19, adjusting for any seasonal or other changes that occur over the same period for those with no COVID-19. Thus, we will refer to this difference-in-differences estimate as the impact of COVID-19 on health care.

We estimated separate impacts for mild COVID-19 and severe COVID-19 using a standard linear regression model with the health care outcome as dependent variable. Independent variables captured that the individual had mild / severe disease vs. no confirmed disease as well as whether the health care service was utilized in the period of question. The interaction between these two categorical variables provided the estimate of interest, i.e. capturing the impact of COVID-19 on the health care outcome in question.

Our model allowed for different estimates in the test week, in the period from the test week through 3 months after test week (the first 1-12 weeks, denoted as 0-3 months,), and from 4 through 6 months (weeks 13 through 24) after test week, using the period from 3 months before and up to the test week as reference. In case age, sex, comorbidities, country of birth or seasonal variations (including policy measures implemented to contain the virus) affected both the severity of COVID-19 and health service utilization [12], we adjusted for the following potential confounders: Age by groups (0-19, 20-39, 40-59, 60-79 and ≥80 years), sex (male/female), comorbidities (categories 0, 1, 2 or 3 or more comorbidities) based on risk conditions for COVID-19 defined by an expert panel [14], birth country (Norway/abroad) and calendar month (12 categories). In this way, we could illustrate the temporal pattern in health care utilization from before to after testing for SARS-CoV-2, for the group with mild COVID-19 compared with the group with no COVID-19, and similarly for the group with severe COVID-19 compared with the group with no COVID-19. We estimated standard errors that allowed for correlation across multiple observations for the same individual over time (clustered standard errors). In addition to providing estimates of the impact of COVID-19 on health care utilization 0-3 and 4-6 months after the test, we also used the model to present plots with rates of utilization for those with no, mild and severe COVID-19 by months (4-weeks periods) before and after testing.

We estimated separate models for all-cause and each of the cause-specific outcomes described in Table 1. To reflect the impact of COVID-19 on long-term health care use, persons who died after the test date were included as non-health care users from the date of death in all our main analyses. Along this line, and to reflect the impact of COVID-19 on long-term overall health, persons who died after the test date were hypothetically included as health care users from the date of death in a sensitivity analysis (all-cause outcomes only). For completeness, we also estimated the proportion of deaths within 3 and 6 months after test date for those with no, mild and severe COVID-19 (95% CIs were calculated based on Wilson). All analyses were run in STATA MP v.16.

Impacts of COVID-19 on all-cause health care utilization

Patients who suffered severe COVID-19 had the highest utilization of primary or specialist health care services in the three months before the test, whereas patients with mild COVID-19 or no COVID-19 had lower and more similar pre-test levels (Table 2). In the three months after the test, utilization was higher in all groups (Table 2). Four to six months after the test, utilization was close to pre-test levels for those who had had mild or severe COVID-19, yet it remained somewhat higher for those with no COVID-19 (Table 2).

Generally, the impact of severe COVID-19 on all-cause health care utilization was higher than the impact of mild COVID-19, and the rates of weekly utilization were increased at 0-3 months post-covid (Table 3, Figure 1). For severe COVID-19 0-3 months after test date, there was a 148% increase in rates of primary care visits, a 76% increase in outpatient visits, and a 472% increase in inpatient visits (Table 3). At 4-6 months after mild and severe COVID-19, we observed little or no increase in all-cause use of primary care or inpatient specialist care (Table 3, Figure 1). However, a ∼24% increase in outpatient specialist care at 4-6 months was observed for severe COVID-19 (Table 3, Figure 1).

• Download figure
• Open in new tab

Figure 1.

Estimated rates of weekly utilization (95% CI) of any, primary, outpatient specialist and inpatient specialist care, from 3 months before to 6 months after week of PCR test for SARS-CoV-2, by severity of COVID-19. Estimates adjusted for age, sex, comorbidities, birth country and calendar month. +/-1 month includes the 1-4 first weeks after/prior to test week, +/-2 months includes the 5-9 weeks after/prior to test week, etc.

Similar patterns were seen in our sensitivity analyses particularly for mild COVID-19, i.e. when we hypothetically treated persons who died after the test date as if they had complaints requiring health care from the date of death. However, the impacts of severe COVID-19 both at 0-3 and 4-6 months after the test date were larger in this analysis (E-Table 1, E-Table 2, E-Figure 1). Even 4-6 months after the test date, the impacts of severe COVID-19 were112%, 268% and 1060% higher on the use of primary care, outpatient- and inpatient specialist care respectively, in the hypothetical case where we set the diseased as having complaints requiring health care (E-Table 2, E-Figure 1).

Altogether, these findings suggest that persons who are mildly affected by COVID-19 have no increase in complaints leading to increased health care utilization beyond 2 months after their test date (Figure 1). The findings also suggest that health care systems are burdened by patients mainly at 0-3 months after testing positive for SARS-CoV-2, with limited persistent increase in health care utilization at 4-6 months.

Impacts of COVID-19 on health care utilization for conditions affecting internal organs

When the cause of health care use was stratified by type of internal organs affected, impacts of COVID-19 on utilization of primary care were in general similar to the results of all-cause utilization (except for group differences in utilization in test week, which were particularly pronounced for respiratory and digestive conditions) (Figure 2, E-Table 3). However, both mild and severe COVID-19 had a large impact on primary care visits due to respiratory conditions (786%-1287% increase) at 0-3 months, but not at 4-6 months (relative change −37%-59% (Figure 2, E-Table 3). Similarly, the impact of severe COVID-19 on utilization of inpatient specialist care due to respiratory conditions at 0-3 months post-test date was increased by 3316% (Figure 2, E-Table 3).

• Download figure
• Open in new tab

Figure 2.

Estimated rates (95% CI) of weekly utilization of primary, outpatient- and inpatient specialist care, for different conditions affecting internal organs from 3 months before to 6 months after week of polymerase chain reaction test for SARS-CoV-2, by severity of COVID-19. Estimates adjusted for age, sex, comorbidities, birth country and calendar month. +/-1 month includes the 1-4 first weeks after/prior to test week, +/-2 months includes the 5-9 weeks after/prior to test week, etc.

The only indications of a long-term impact of COVID-19 were a 166%-337% increased rate of visiting outpatient specialist care for circulatory and respiratory conditions both at 0-3 and at 4-6 months post-test (Figure 2, E-Table 4). In addition to an expected long-term circulatory and respiratory dysfunction among persons having had severe COVID-19, our estimates for outpatient- and inpatient specialist care for persons having undergone severe COVID-19 also imply dysfunction of the endocrine, metabolic and nutritional systems. More specifically, we observed a 168% and 791% increased rate in outpatient- and inpatient specialist care visits (respectively) due to endocrine, metabolic and nutritional conditions at 0-3 months after, but not 4-6 months after, the test date (Figure 2, E-Table 4, E-Table 5). COVID-19 patients had no subsequent increased health care use for digestive or genitourinary conditions (Figure 2, E-Tables 3-5).

Impacts of COVID-19 on health care utilization for conditions affecting external, sensory and other organs

We observed no impact of COVID-19 on utilization of primary, outpatient- or inpatient specialist care for eye and ear-, musculoskeletal-, mental-, skin- or blood conditions (Figure 3, E-Table 6-8). However, both mild and severe COVID-19 had a small impact on primary care visits due to general and unspecified conditions (4-7% increase at 0-3 months, no increase for 4-6 months post test) (Figure 3, E-Table 6). Severe COVID-19 also had a large impact on outpatient specialist care at 0-3 months (48% increase) and at 4-6 months (40% increase) due to general and unspecified conditions, as well as on inpatient specialist care at 0-3 months (431% increase) after being tested (Figure 3, E-Table 7, E-Table 8).

• Download figure
• Open in new tab

Figure 3.

Estimated rates (95% CI) of weekly utilization of primary, outpatient- and inpatient specialist care, for different conditions affecting external, sensory and other organs from 3 months before to 6 months after week of polymerase chain reaction test for SARS-CoV-2, by severity of COVID-19. Estimates adjusted for age, sex, comorbidities, birth country and calendar month. +/-1 month includes the 1-4 first weeks after/prior to test week, +/-2 months includes the 5-9 weeks after/prior to test week, etc.

Further, although no group differences could be observed that implied a relative increase in health care use for mental conditions, it should be noted that there was a large increase in the use of both primary and specialist care for mental causes for all groups, i.e. both among those without COVID-19 and those with mild or severe COVID-19 (Figure 3).

Principal findings

In this population-wide study of 1,257,831 persons comprising everyone tested for COVID-19 in Norway by November 1^st 2020, we found no impact of mild COVID-19 on deteriorated health or higher health care use that persisted beyond 2 months after having tested positive. We also found that severe COVID-19 increased the health care use after 0-3 months, particularly for respiratory, circulatory, endocrine/metabolic/nutritional and general/unspecified conditions. Circulatory and respiratory conditions as well as general/unspecified conditions were the only potential post-covid conditions giving increased long-term health care use at 4-6 months post-covid, however only for those who had undergone severe COVID-19.

To our knowledge, this study is the first to compare long-term impacts for persons with mild and severe COVID-19 and to study whether their health care use is increased relative to persons with no confirmed COVID-19. With our prospective design, we have also for the first time shown how COVID-19 patients are affected and when they use which type of health care, i.e. we could study both primary care, outpatient- and inpatient specialist care use for any condition affecting internal-, external-, sensory- and other organs. Thus, we provide a comprehensive picture of the overcoupling post-covid health and health care use. In addition to the complete data including everyone tested for COVID-19, other strengths of our study are the methods used – we could take account of the pre-covid health care use for an effective comparison of post-covid health care use. By studying time as relative from the week of COVID-19 testing, and comparing with everyone who tested negative for COVID-19, we could adjust for calendar month and in this way also adjust for seasonal variations in both transmission and health care use. This attractive feature of the difference-in-differences model becomes particularly evident for our findings about utilization of health care for mental conditions (Figure 3). Utilization for mental conditions did increase for patients with COVID-19, both mild and severe, from before to after the test week, indeed it more than doubled from 3 months before the test to 6 months after (Figure 3). However, persons with no COVID-19 suffered a similar or even more pronounced increase in mental conditions over the same period (Figure 3), which suggests that it is the pandemic and its countermeasures – not actually undergoing COVID-19 – that induces mental distress. The difference-in-differences model accounts for this [12], and thus provides no estimates of positive impacts of COVID-19 on health care utilization for mental conditions (E-Table 6, E-Table 7).

Comparison with related studies

Our study confirms and sheds new light on recent studies of post-covid complaints and health care use among patients who had suffered severe COVID-19. More specifically, we confirm findings from a matched case-control study reporting increased hospital re-admissions due to pulmonary pathophysiology, diabetes and cardiovascular disease among brits who had been hospitalized for COVID-19 [7]. Here we could provide a time frame for the additional health care use among COVID-19 patients, showing that the largest increase due to such conditions comes 0-3 months post-covid for most causes. Yet, there is still an increase in outpatient specialist care visits due to respiratory and circulatory conditions at 4-6 months post-covid. Remarkably, both outpatient and inpatient health care use for circulatory, respiratory, endocrine/metabolic/nutritional conditions as well as general and unspecified conditions, were greatly increased after COVID-19 among persons having undergone severe disease. The impact on outpatient and inpatient visits for such conditions increased between 48%-3316% at 0-3 months and 40%-246% at 4-6 months. We suggest the link between severe COVID-19, multiorgan dysfunction including general and unspecified symptoms at 0-6 months particularly after severe disease, as a topic for further study.

While the long-term deterioration in health for patients who were hospitalized with COVID-19 has been documented in several studies [2-4, 6], evidence on so-called “post-covid syndrome”, “post-acute covid-19” or “long covid” is very limited for the vast majority of persons that have a mild disease course of COVID-19 [15]. This syndrome has been described as a multisystem disease (based on patient-reported data), sometimes occurring after a relatively mild acute illness, i.e. extending beyond three or twelve weeks after the onset of first symptoms [9]. Symptoms reported are cough, low grade fever, and fatigue, shortness of breath, chest pain, headaches, neurocognitive difficulties, muscle pains and weakness, gastrointestinal upset, skin rashes, metabolic disruption, thromboembolic conditions, and mental health conditions [9]. We study every resident in Norway with detected mild COVID-19, and find no evidence of persisting complaints causing a need for health care beyond 2 months after having tested positive (Figure 1). However, possibly worth noting is that we observed a minor (4%) increase in visits due to “general and unspecified conditions” in primary care at 0-3 months after positive test. Such general and unspecified conditions may include a wide range of symptoms and clinical signs not classifiable or detectable on imaging or laboratory tests, such as ICPC-2 codes “A01 – generalized pain”, “A02 - chills”, “A05 – feeling ill” and “A04 - fatigue”. Given previous studies [2-4, 9], the minor increase in general and unspecified conditions in primary care deserves attention in future research despite no indications of overall elevated health care use 4-6 months after mild COVID-19 in our population-wide analyses.

Interpretation

Important to the interpretation of our findings are the high rate of health care use during the test week, particularly for primary care and inpatient specialist care. These peaks are expected, as we also include digital GP consultations (about 50% of all consultations) and routine PCR testing related to hospital care. For example, digital GP consultations are likely to increase when COVID-19 is suspected and patients are quarantined, and many patients who are scheduled for elective hospital care are likely to be tested by routine to avoid transmission in hospitals. Along this line, parts of the increase in health care use during our post-test period may be explained by a routine follow-up visit that is recommended for persons having undergone COVID-19 [16]. Further, we had no cut-off for when COVID-19 started or ended, which may explain the increased health care use at 0-3 months but not at 4-6 months. Considering patient-reported descriptions of COVID-19 feeling like a heavy and long-lasting flu that has different durations for different patients [17], we chose not to set a cut-off for the end of the disease. Rather, we focused on the test date and could shed light on the time frame for the need for health care also when the infection was still ongoing. In that regard, COVID-19 related deaths may explain the high need for health care at 0-3 months, which was followed by a drop in health care use at 4-6 months (12% of patients with severe COVID-19 had died within 3 months after the positive test). These findings from our main analyses may therefore reflect the burden of COVID-19 patients to the health services, accounting for the trivial fact that patients who die will not enter health services. However, to shed light on the impacts of COVID-19 on the health of the infected patients, in our sensitivity analysis, we hypothetically treated all individuals who died during the post-test period as if they used health care services from their date of death. This hypothetical analysis may be more informative for understanding the patient-relevant burden of COVID-19. Not surprisingly, this hypothetical analysis reveals a large impact of severe COVID-19 on hypothetical health care use, while the impact of mild COVID-19 remains negligible beyond three months after the test for SARS-CoV-2 (E-Figure 1, E-Table2).

Of further relevance to the interpretation of findings is that we also find that mild COVID-19 may lead to a reduced health care use after 4-6 months (i.e. negative effect estimates), which cannot be attributed to deaths. More specifically, the utilization of outpatient- and inpatient specialist care does slightly increase from 3-0 months before, to 4-6 months after the test week for those with mild COVID-19, while the utilization is virtually unchanged for primary care (Table 2, Figure 1). Explanations for these findings are unknown. However, although it seems reasonable that mental distress related to the pandemic and its countermeasures affects persons with and without COVID-19 to a similar extent (as observed in Figure 3), the initial decline and subsequent recovery of utilization of health care services across groups of the general population [18], deserve further attention in future research.

Potential limitations

Some important limitations should be mentioned. First, our organ-wise studies of primary care, outpatient- and inpatient specialist care might include different diagnoses due to the different setups of the ICPC-2 and the ICD-10 system. As an example, the ICPC-2 chapters include separate codes for patient-reported symptoms and signs in addition to doctors’ diagnoses, whereas the ICD-10 chapters mainly include diagnostic codes that are based on clinical and laboratory tests or imaging. Along this line, the ICPC-2 system includes diagnostic codes for cancers in the ICPC-2 chapters that were included in this study, whereas ICD-10 diagnostic codes for cancers have an own chapter that was not included in our study. However, because our main aim was to provide a broad overview of health and health care use after mild and severe COVID-19, and because we could contrast with pre-test patterns and patterns for those with no COVID-19 in the difference-in-differences model, we regard this to be of small relevance to the interpretation of our findings.

A second limitation may be the limited test capacity in the beginning of the pandemic. For this reason, we might have missed a large part of the earliest mild COVID-19 cases. Still, only persons with confirmed negative test were included in the comparison group and we expect no over- or underestimation of results for the group having mild COVID-19. It should further be noted that although we defined persons having a positive PCR test as if they had mild COVID-19, these persons could be free of symptoms. We also chose to set the first negative test as the time point for comparison for those with only negative tests. However, if we instead used the last negative test as the time point, we would have had a shorter follow-up time, because a large number of people who test themselves several times. Another alternative would be to randomly choose one test date from all the negative tests, yet we believe this would have led to overrepresentation of observations in periods with massive testing [19]. In an attempt to keep our analyses as transparent as possible, we also refrained from applying matching techniques to construct the comparison group with no COVID-19.

A third limitation may be that we lack data from some private rehabilitation institutions in the BeredtC19-register. Public hospitals will often have their own department for rehabilitation, yet most rehabilitation treatment occurs at private institutions that were not included in our data sources. Again, because we could contrast with pre-test patterns and patterns for those with no COVID-19 in the difference-in-differences model [12], we regard this to be of small relevance to the interpretation of our findings.

Finally, our analyses are obviously based on observational data, and what we refer to as impacts of COVID-19 on health care utilization may of course be related to a number of confounders. While the difference-in-differences model accounts well for time invariant individual characteristics, as well as time varying health care utilization that also affects persons with no COVID-19, we cannot be certain that the temporal pattern in utilization of those with no COVID-19 is a reasonable counterfactual for the health care utilization of patients with COVID-19. We do observe, though, that the utilization rate for health care in the months before the test week is almost identical in the group with no and mild COVID-19 (Figure 1), which is what we would have expected if mild COVID-19 had been randomly attributed in the test week.