Introduction
Since the outbreak of the Coronavirus disease 2019 (COVID-19), there has been a growing concern regarding the effect of the pandemic on public mental health [
1]. Governments worldwide have applied various containment measures to prevent the spread of the virus, which could negatively affect individual’s well-being and mental health [
2]. Many individuals may have experienced increased stress, loneliness, and reduced social contact, which are well-known risk factors for developing mental disorders [
3]. For instance, Robillard et al. [
4] found an increase in the proportion of Canadian respondents without a history of mental disorders who screened positive for depression and anxiety during the pandemic. On the other hand, results from systematic reviews show that an initial increase in mental health problems during the early stages of the pandemic was followed by a decline during later stages, which indicates a high level of resilience in the population [
5,
6]. Qualitative studies reveal that some individuals experienced the shift to home office and cessation of commuting created time to do more positive activities, such as spending more time with loved ones, exercising and parenting [
7]. Some individuals with mental health problems reported that their life became better during the pandemic, because of a slower pace and less pressure to perform in the daily life [
8]. For instance, the pandemic created a more flexible calendar, less meetings, better circadian rhythm and sleep, better family relations and more time to make decisions in everyday life. Many individuals therefore seem to either have been mostly unaffected by the pandemic or were even doing better during the pandemic [
9]. This could be due to reduced risk factors and less social demands, caused by the confinement and social distancing measures. For instance, people with social anxiety may have experienced decreased stress due to social distancing measures, such as mandatory home office, homeschooling, and fewer social events.
How individuals react to challenging events is based on their personal resources. It is therefore likely that the COVID-19 pandemic impacted individuals differently based on their circumstances. For instance, younger people and people living alone have experienced more loneliness than older adults [
10]. Parents may have been more affected by lockdown than adults without children, as parents have experienced increased daily hassles due to the additional stress of trying to combine home office and homeschooling [
11]. Whether the pandemic has affected the relative risk of individuals being diagnosed for the first time with a mental disorder is uncertain.
Studies across several countries have investigated and compared incidence rates of specific mental disorders before and during the pandemic, but the results are inconsistent. A study from the USA used electronic health records to investigate the incidences of diagnosed eating disorders among children and young adults (0–30 years) during the first year of the pandemic [
12]. Compared to 2019, the incidence of eating disorders increased in 2020, particularly among young (aged 10–19 years) females. In Lithuania, a study using data from a National health care registry found that the incidence rates for post-traumatic stress disorder (PTSD), adjustment disorder, and major depressive disorder (MDD) decreased in 2020 compared to 2018 and 2019 [
13]. In line with this finding, a study from the UK also found a reduction in the incidence of primary care-recorded anxiety and depression in April 2020 compared to 2019, and the decrease was most pronounced among adults (18–64 years) [
14]. By September 2020, incidence rates of anxiety and depression were similar to expected levels in England but remained lower than expected in Northern Ireland, Scotland, and Wales [
14].
The previous studies have some important limitations. For instance, the incidence rate of eating disorders was only investigated among children and young adults [
12], and the incidence rates of PTSD, adjustment disorder, and MDD was examined using data from an insurance registry [
13]. In addition, all studies mentioned above have only used data from the first year of the pandemic [
12‐
14].
There are many possible stressors connected to the COVID-19 pandemic, such as fear of getting infected or infecting others, grief of losing a close friend or relative, financial insecurities due to job loss or temporarily being laid off, and social isolation. Such stressors could increase mental health problems, and lead to an exacerbation or development of new stress-related disorders, such as anxiety, depression, and PTSD [
15,
16]. In addition, quarantine, isolation, a sense of loss of control, and changes in routines could have a negative impact on eating behavior [
17,
18], and increase symptoms or lead to the development of new eating disorders in those at risk [
19].
We extend previous research by using nationwide registry data capturing all contacts with primary- and specialist health care services and examining incidence rates beyond the first year of the pandemic. Based on the above considerations, we aimed to investigate if the incidence rates of mental disorders hypothesized to be influenced by the pandemic-related stress and reduced social contact (i.e., MDD, anxiety disorders, obsessive–compulsive disorder (OCD), adjustment disorder, and eating disorders) changed during the pandemic (2020–2021) compared to pre-pandemic years (2015–2019) in primary- or specialist health care.
Discussion
Our results indicate that the COVID-19 pandemic has affected the incidence of mental disorders differently. The incidence rates of OCD in specialist health care and eating disorders in both primary- and specialist health care increased during the pandemic. The increase was only apparent among women and most pronounced in the two youngest age groups. On the other hand, in some age groups, the incidence rates of primary care-recorded depression and phobia/OCD, and specialist care-recorded phobic anxiety disorders decreased during the pandemic compared to the pre-pandemic years. There were no differences in the primary care-recorded incidence rates of anxiety disorder and PTSD and in the specialist care-recorded incidence rates of depression, other anxiety disorders, and adjustment disorders during the pandemic compared to the pre-pandemic years. The pandemic affected incidence rates of some disorders and in specific age groups, while most incidence rates followed the underlying trend from before the pandemic.
We found that primary care-recorded incidence rates of depression among men in all age groups and among women in the oldest age groups were lower than predicted. This decrease was not found in specialist health care. The result is surprising, as systematic reviews and meta-analyses have found a higher prevalence of self-reported depression in the general population during the pandemic [
22‐
24]. However, the reduction in incidence rates are in line with both Carr et al. [
14] and Kazlauskas et al. [
13]. Our finding extends previous results as we also found a lower incidence rate of depression in primary health care during the second pandemic year (2021).
Further, we found decreased incidence rates of phobia/OCD among men in the oldest age group and women in the youngest and oldest age groups in KUHR, and decreased incidence rates of phobic anxiety disorders among men in all age groups and women aged 25–39 years in NPR during both pandemic years. Conversely, we found an increase in the incidence rate of specialist care-recorded OCD in 2021 among women in the youngest age group, which was not found among men. A decline in the incidence rates of phobia/OCD in primary health care and phobic anxiety disorders in specialist health care is in contrast to the prevalence literature using survey data, as some studies found elevated symptoms of phobic anxiety during quarantine [
25] and lockdown [
26]. However, increased incidence rates of OCD in specialist health care align with systematic reviews that have found an increased prevalence of OCD symptoms and the emergence of new symptoms [
27‐
29].
There might be several explanations for the finding of decreased incidence rates of depression and phobia/OCD in primary health care, and phobic anxiety disorders in specialist health care during the pandemic. Firstly, the social distancing measures during the pandemic might have reduced some individuals' mental distress [
13,
14]. People were encouraged, and in some periods forced, to have home office and homeschooling. Most social gatherings were canceled or reduced to a limited number of participants. Consequently, some people might have avoided situations that could trigger symptoms of mental disorders. This could have reduced the number of people receiving their first-time diagnosis of a mental disorder. Secondly, access to mental health care and GPs in Norway was reduced during periods with strict social distancing measures [
30,
31]. For instance, people with newly arisen respiratory tract symptoms and those suspected infected with or exposed to the SARS-CoV-2 virus were advised not to have a face-to-face consultation with their GP or psychologist [
30‐
33]. This might have discouraged some individuals from seeking help. Thirdly, some individuals might have feared becoming infected at the GPs office or mental health services and therefore stayed home. Since the incidence rates of depression, phobia/OCD, and phobic anxiety disorders have been either stable or had an increasing trend before the pandemic, this might indicate that some people needing mental health care did not seek help during the pandemic. If so, this is alarming as untreated mental disorders have been associated with a higher risk of suicide ideation, financial problems, family problems, and discontinuation of work and higher education [
34]. Lastly, the results could be explained by true decreases or delayed help-seeking. If the results are due to delayed help-seeking, this might lead to higher incidence rates in the post-pandemic period. Future studies should investigate incidence rates across a more extended time after the pandemic to see whether the rates will rebound in the coming years.
Lastly, we found that the incidence rates of primary- and specialist care-recorded eating disorders increased considerably during the pandemic among women. The increase was apparent across all age groups, except those aged 25–39 years in KUHR, but was most prominent among the youngest age group in 2021. In primary- and specialist health care, the observed incidence rate among 18–24-year-olds in 2021 was more than 40% higher than predicted. Several countries have reported an alarming increase in the prevalence of eating disorders, especially among young people [
35]. Our result aligns with Taquet et al. [
12], who found increased incidence rates among children and young adults (0–30 years) in 2020. However, they found the largest increase among girls aged 10–14 and 15–19. Increased use of primary- and specialist health care for eating disorders during the pandemic have previously been documented among Norwegian children and adolescents [
36]. Our finding extends previous results, as we found an increase also among adult women and during the second pandemic year (2021).
The findings of increased incidence rates of OCD and eating disorders among women during the pandemic, especially in the youngest age group, might be explained in several ways. The general finding of increased incidence rates among women could reflect that women have been more vulnerable during the pandemic [
37], as systematic reviews have found more mental health problems among this group compared to men and other groups [
6,
38,
39]. The finding of a larger increase in incidence rates among those aged 18–24 years aligns with studies documenting that young adults might have been more affected by changes in everyday routines, strict social distancing measures, and closures of university campuses leading to homeschooling [
40]. This could explain why prevalence studies have demonstrated increased symptoms of mental distress among this group during the pandemic [
41,
42]. In addition, young adulthood is a period where the emergence of psychiatric disorders usually takes place [
43]. For instance, OCD is more common in adolescence and adulthood among females compared to males [
44]. In addition, an increasing trend of eating disorders, especially anorexia nervosa, has been reported in individuals aged 15–19 years, and this trend is found among women and not men [
45]. Taken together, our results might indicate that the pandemic and its related consequences put even more young women at risk of developing OCD and eating disorders.
Regarding the increased incidence rate of OCD in 2021 among women aged 18–24, there was much focus on disinfection, cleaning, and personal hygiene to help prevent the spread of the virus. There was also much uncertainty regarding how dangerous and contagious the virus was, which resulted in a massive fear of getting infected and infecting others. This might have triggered OCD symptoms, especially among those with contamination-related symptoms, and may have led to more people being diagnosed with OCD for the first time [
29]. It is uncertain how the pandemic has affected specific subtypes of OCD, as most research has focused on contamination-related OCD symptoms [
28]. Our study also shows that there has been a somewhat increasing trend in the incidence rates of OCD among women in Norway during the study period, which could partly explain the result.
Regarding eating disorders, many factors could contribute to the increased incidence rates among women during the pandemic. Firstly, almost everyone experienced disruptions in their daily routines, such as periods with closed gyms and other indoor activities, which could increase concerns regarding weight and appearance [
46]. In addition, social media, either in the form of harmful eating/appearance-related content or stressful and traumatic news articles related to the pandemic, could also trigger symptoms of eating disorders [
46]. In general, the COVID-19 pandemic may have acted as a global stressor that triggered the development of OCD and eating disorders among vulnerable individuals.
This study has several strengths. Firstly, we used nationwide data on incidence rates from both primary- and specialist health care, thus covering Norway's entire public health care system. Secondly, our data captured all patient encounters between 2006/2008–2021 from the entire population, thus making selection bias unlikely. Thirdly, with access to incidence rates from both 2020 and 2021, we added to and extended previous research, which only investigated incidence rates during 2020, the initial stages of the pandemic. Fourthly, we rely on diagnostic codes for mental disorders diagnosed by qualified health professionals in primary- and specialist health care. The diagnostic codes are classified according to two international classification systems (ICPC-2 and ICD-10), which ensures an international standard for diagnosing health information. One Norwegian study examined the correspondence between mental disorder diagnoses based on structured diagnostic interviews (Composite International Diagnostic Interview; CIDI), and diagnoses set in primary- (KUHR) and specialist health care (NPR), and found that diagnoses in health registries have excellent specificity because of the few false positives, and moderate sensitivity [
47].
Our study also has some limitations. Firstly, we investigated incidence rates in the two health registries separately. Consequently, an individual could have received a first-time diagnosis of depression in specialist health care in 2016 but have been registered with a first-time diagnosis of depression in primary health care in 2015. This could lead to a higher number of incident cases, as the same individual may be counted as a unique case in the two separate registries. Secondly, subject-level data are available from 2006 in KUHR and from 2008 in NPR. Data during the first years will not only capture true incident cases, but also prevalent cases, as individuals might have been in contact with primary- and specialist health care before the inception of the registries. Therefore, we investigated incidence rates between 2015–2021. However, this resulted in a short reference period before the pandemic outbreak (5 years), reducing the prediction models' statistical power to detect significantly changes in incidence rates during the pandemic years. Thirdly, health professionals in primary- and specialist health care might have different diagnostic practices. In primary health care, there are no need for a referral, leading to more patients presenting with a wide range of health complaints. Hence, the diagnostic practice might be more affected by pragmatic issues, such as patients requiring sick leave and different interpretations of subjective health complaints, which might increase the number of reported cases [
48]. In specialist health care, patients have been referred based on having a moderate or severe mental disorder. Due to a more selected population and fewer patients who are in contact with specialist health care, this might lead to an underrepresentation of true cases [
47]. However, we have no reason to believe that the putative differences in diagnostic practices in primary- or specialist health care changed during the pandemic compared to pre-pandemic years. Fourthly, in ICPC-2, phobia and OCD are merged in the same diagnostic code (P79). This makes it impossible to interpret whether the reduced incidence rates of P79 were due to changes in one or both disorders. The finding of decreased incidence rates of phobic anxiety disorders (F40) and increased incidence rates of OCD (F42) in specialist health care using ICD-10, could indicate a corresponding decrease of phobia and increase of OCD in primary health care. Lastly, the present study did not adjust for multiple comparisons, which could increase the likelihood of making a Type I error. This may affect the interpretation of results, as some of the significant findings might have occurred by chance. However, multiple comparison correction could increase the likelihood of making a Type II error, rendering the results more challenging to interpret. By presenting all
p-values and PIs for all tests and including the descriptive trends in Supplementary Material
1, we are transparent regarding our results. This makes it possible for the readers to assess the plausibility of our findings.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.