Incidence and Determinants of COVID-19 in Patients Seeking Treatment for Substance Use Disorder: A Patient-Based Linkage Study

The first case of COVID-19 was reported in China in December 2019 [1]. Just over 1 month later, the first case was reported in Spain, and the first case in the autonomous region of Catalonia was reported almost 2 months later, on 25 February 2020 [2, 3]. Spain was one of the countries hardest hit by COVID-19 pandemic [2, 4] particularly during the first wave, despite being one of the European countries which adopted the strictest protection measures. By December 2021, Spain had reported 5,422,168 cases and over 18% of the reported cases occurred in Catalonia, making it the autonomous region with the highest number of reported cases [5].

Studies done in Spain and Catalonia, looking at the first wave of the pandemic, showed suffering one or more comorbidities such as, chronic kidney disease, chronic obstructive pulmonary disease, heart disease, hyperlipidaemia, hypertension, cancer, obesity, and type 2 diabetes mellitus are associated with a higher risk of COVID-19 infection and poorer outcomes (hospitalizations, admission to intensive care, and deaths) [2, 6]. People with substance use disorder (SUD) are more likely to suffer cardiac, pulmonary, metabolic, and immune diseases as a result of chronic alcohol or other substance use [7‒9]. Therefore, they are more vulnerable to COVID-19 infection and have a worse prognosis [10‒12]. This was further confirmed by a study in the USA of over 73,000,000 unique patient electronic health records which found that those with SUD, especially those diagnosed with opioid, alcohol, cocaine, or tobacco abuse had a greater risk of COVID-19 infection and COVID-19-related hospitalization and death, compared to those without SUD [13]. In addition to comorbidities and worse clinical outcomes, people with SUD have worse socioeconomic conditions, including homelessness, imprisonment, and more risky behaviours such as sharing drug using equipment [12]. This group also live with the effects of stigma and face more barriers to accessing the health system [14], which can in turn, create a greater risk of exposure to COVID-19 and to suffering more clinical complications.

Nonetheless, in spite of being at greater risk of infection [12, 15], little is known about the cumulative incidence of COVID-19 in people with SUD or of the social and clinical determinants of infection. This study aims to: (1) determine the cumulative incidence of COVID-19 from 25 February to 31 December 2020, among people who sought treatment in the Catalan public health system between 1 January 2018 and 31 December 2019, for abuse of the following substances: alcohol, cocaine, cannabis, and heroin; (2) determine the sociodemographic features, e.g., sex, age, socioeconomic position (SEP), country of birth, type of residence, substance of use, and clinical features, such as HIV and adjusted morbidity groups (AMG), associated with COVID-19 infection in people with SUD; (3) compare the cumulative incidence of COVID-19 infection between people with SUD and the general population during the study period.

The patient-based retrospective observational cohort study was conducted.

In the 63 Drug Dependence Centres (CAS, Catalan acronym) of the Catalan Health Department, there were 26,773 initiations to SUD treatment between 1 January 2018 and 1 January 2019 (13,075 in 2018; 13,698 in 2019). Patients who died before February 25,2020 and those without an anonymous health identifier were excluded from the study (294 and 668 patients, respectively). Additionally, in the case of a person entering treatment more than once during 2018–2019 only the first instance was included (1,476 treatments excluded). A further 1,243 treatments for substances other than heroin, cocaine, cannabis, or alcohol were excluded because they were a made up of a heterogeneous set of substances (see online suppl. annex 1; for all online suppl. material, see https://doi.org/10.1159/000528647).

Therefore, the final cohort comprised 23,092 people who sought treatment for abuse of, or dependence on, alcohol, cocaine, cannabis or heroin between 1 January 2018 and 31 December 2019. These substances represent 94.89% of the total number of treatment initiations.

Information was obtained from three sources: first, information regarding initiation to treatment was obtained from the Catalan drug dependence information system (SIDC). Since 1987, professionals working in the CAS have notified initiations to treatment via the SIDC, including sociodemographic characteristics (sex, age, country origin, and type of residence) and the substance for which treatment was sought. The second source was the administrative database of the Catalan Health Surveillance System (CHSS). This was accessed via the Public Program for Data Analysis for Research and Health Innovation (PADRIS). The CHSS includes (among other datasets): the Central Registry of Insured People (RCA, Catalan acronym), the Registry of the Minimum Basic Dataset (CMBD, Catalan acronym), the Registry of Adjusted Morbidity Groups (AMG), and the COVID-19 Epidemiological Surveillance Registry (RSA-COVID-19). The third source was the longitudinal Catalan PISCIS study cohort of people with HIV [16].

The RCA collects information on employment status, annual income, and benefits received from the social security system. These variables allow calculation of the pharmaceutical co-payment, and this was used as a proxy for SEP, as in previous studies [17, 18].

HIV status was obtained from the CMBD, an administrative register which collects information on diagnoses made in the Catalan health centres and coded according to the International Classification of Diseases-10 (ICD-10). The Emergency CMBD and the Hospital and Primary Care CMBD were used. Information on HIV infection was complemented by information from the PISCIS cohort.

The AMG is a measure of morbidity adapted to the Catalan health system, and validated by the World Health Organization, which allows for stratifying the population by morbidity and clinical complexity [19]. In addition, it produces less variability and greater precision than other risk tools, such as the clinical risk groups [20] and consists of assigning each individual a numerical value, with a higher value indicating greater clinical complexity.

Data on COVID-19 infection for SUD patients were obtained via the RSA-COVID-19, which contains the results of diagnostic tests done in the Catalan health system of all suspected cases. According to the protocols of the Spanish Health Ministry, a suspected case was defined as those people who present a clinical profile of acute respiratory infection or other symptoms including: loss of smell and/or taste, diarrhoea, headache and muscular pain; or close contacts of a positive case [21]. Data on the cumulative incidence of COVID-19 infection in the general population were obtained from the Catalan Health Department Register of COVID-19 cases. Clinical and sociodemographic data obtained from the CHSS were automatically combined with the data on initiations to treatment from the SIDC using an individual anonymous health identifier.

Sociodemographic variables included were sex; age (≤30 years, 31–40 years, 41–50 years, ≥51 years); country of origin (born in Spain or another country); SEP (very low: receiving benefits and/or unemployed; low: annual income <EUR 18 000; mid-high: annual income >EUR 18 000); type of residence (stable, homeless, institutionalized). People living in shelters or in prison were included in the category “institutionalized.”

The substance-related variable considered was principal substance for which treatment was sought (alcohol, cocaine, cannabis, or heroin). Clinical variables considered were diagnosis of HIV according to CMBD and/or the PISCIS cohort; and AMG, with the population classified by four percentile categories (P50 baseline risk, P75 low risk, P85 moderate risk, P90 high risk).

A person was considered to have been infected by COVID-19 if there was a confirmed positive polymerase chain reaction (PCR) result recorded in the RSA-COVID-19 between 25 February 2020 and 31 December 2020. It was assumed that no infection had occurred if a positive PCR result did not appear.

A descriptive analysis was undertaken of the study population’s sociodemographic and clinical features, stratified by type of substance for which treatment was initiated. The descriptive results were presented by absolute and relative frequency, and differences in distribution patterns by substance were assessed by χ² test for categorical variables and Kruskal-Wallis test for continuous variable. p values were included with a significance level of 5%.

Bivariate analysis was conducted to determine differences between SUD patients with COVID-19 infection and those without infection, according to: substance, sex, age, SEP, country of birth, type of residence, HIV diagnosis, and AMG. Pearson’s χ² test for categorical variables and Wilcoxon rank-sum for continuous variable were carried out and p values were included with a significance level of 5%. The cumulative incidence of COVID-19 was calculated from 25 February 2020 to 31 December 2020, and stratified by sex and substance type.

Finally, crude and adjusted risk factors for COVID-19 infection were analysed. Adjusted risk factors were calculated using a log-link binomial generalized linear model [22], estimating relative risk (RR), confidence intervals (CIs), and p values. Variables included in the final model were established using the Wald test, considering the Akaike information criterion and scientific evidence of factors associated with COVID-19 infection. RR was estimated with their respective 95% CIs and associated p values. Statistical analysis was undertaken using STATA version 15.

Between 2018 and 2019, in Catalonia, 23,092 people initiated treatment for substance abuse of or dependence on heroin, cocaine, cannabis, or alcohol. Of these, 47.98% sought treatment for alcohol abuse or dependence, 27.46% for cocaine, 13.61% for cannabis, and 10.95% for heroin.

Table 1 shows the sociodemographic and clinical characteristics of the 23,092 people who initiated treatment, stratified by principal substance. Statistically significant differences were observed for sex, age, country of origin, SEP, type of residence, HIV diagnosis, and AMG. For all substances, treatment seeking was higher among men. For women, the largest proportion was seen for alcohol (26.20% of those seeking treatment) and the lowest for heroin (11.72% of those seeking treatment).

Those who commenced treatment for alcohol were older, with a median age of 46.86 years (IQR = 38.86–55.16). Among those seeking treatment for alcohol, 11.69% were in a high-risk AMG. Of treatment initiations for heroin, 32.98% of people had a very low SEP, 26.34% were migrants, and 17.54% had HIV infection.

During 2020, of the 23,092 people who made up the study population, 8,810 (38.15%) appeared in the RSA-COVID-19 as suspected cases. Among the suspected cases, there was little variation by substance (heroin: 34.09%; cannabis: 36.52%; cocaine: 39.00%; alcohol: 39.05%). There were a total of 601 cases with a positive PCR test result for COVID-19 (Fig. 1), representing a cumulative incidence during 2020 of 2.60% (95% CI = 2.40–2.81%) in the population of people seeking treatment for substance use.

Table 2 shows the cumulative incidence of COVID-19 by sex and substance type. A significantly higher cumulative incidence was observed among people in treatment for alcohol abuse or dependence (3% [95% CI = 2.70–3.34]) compared to those treated for heroin (1.94% [95% CI = 1.47–2.56]). There was no statistically significant difference between men and women in the cumulative incidence of COVID-19 infection. Figure 2 shows the cumulative incidence of COVID-19 from 25 February to 31 December of 2020, among SUD patients by type of substance. Meanwhile, according to the Catalan Health Department Register of COVID-19 cases the cumulative incidence of COVID-19 infection in the general population was 3.86% (95% CI = 3.85–3.87).

Of the 601 people who tested positive for COVID-19, 55.41% (n = 333) had sought treatment for alcohol use, and only 8.15% (n = 49) for heroin. Those infected with COVID-19 were significantly older: 31.28% were older than 51 years, compared to 24.48% of those not infected with COVID-19. There was also a significantly higher proportion of migrants who tested positive for COVID-19 (25.48% positive vs. 15.35% not positive) and of people living in institutionalized residences (9.64% positive vs. 7.47% not positive). Regarding the clinical variables, among those positive for COVID-19 there was a greater proportion of people with HIV (6.32% positive COVID-19 vs. 3.92% not positive COVID-19) and people with a high-risk AMG index (15.21% positive COVID-19 vs. 7.90% not positive COVID-19) (see Table 3 ). Significant differences were not seen by sex or socioeconomic level.

Table 4 shows crude and adjusted RR for COVID-19 infection. The adjusted RR shows higher risk of COVID-19 infection among those who sought treatment for alcohol abuse or dependence (ARR = 1.62 [95% CI = 1.14–2.31]) compared with those treated for heroin; higher risk for migrants compared to those born in Spain (ARR = 2.12 [95% CI = 1.74–2.58]); and higher risk for people living in institutionalized residences compared to homeless people (ARR = 2.09 [95% CI = 1.25–3.49]).

Regarding clinical variables, higher risk of COVID-19 infection was observed for people with HIV (ARR = 1.59 [95% CI = 1.10–2.30]) and among those with high AMG (ARR = 1.84 [95% CI = 1.28–2.65]) compared to baseline AMG. No significant differences were seen by sex or SEP.

This patient-based study of people with SUD shows that the cumulative incidence of COVID-19 in more than 23,000 patient initiations to treatment for alcohol, cocaine, cannabis, or heroin was 2.60% (95% CI = 2.41–2.82). Among those who sought treatment for alcohol abuse or dependence in the 63 CAS, the cumulative incidence was significantly higher than for those who sought treatment for heroin. In addition, being a migrant, living in an institutionalized residence, having HIV, and high morbidity according to AMG, was associated with COVID-19 infection in people who sought treatment for SUD in 2018 and 2019. Moreover, despite SUD patients’ clinical vulnerability and their level of social disadvantage, the cumulative incidence of COVID-19 was lower in SUD patients than in the general population of Catalonia (3.86% [95% CI = 3.85–3.87)]), during the same period.

It should be mentioned that PCR testing for COVID-19 was not routinely available in Spain until the first quarter of 2020, which may have led to an underdiagnosis of infections in both, the general and SUD populations. Moreover, SUD patients have to overcome more barriers to accessing the health system due to the stigma [23], which could mean less screening and diagnosis of the infection among this group. Among SUD patients, there may also be differences in stigma depending on the type of substance used. Analysis of the sociodemographic characteristics of the study population showed patients treated for heroin had a higher prevalence of homelessness, very low SEP and were more likely to be migrants than other SUD patients. These characteristics contribute to the fact that patients who use heroin suffer greater social isolation and stigmatization than other SUD patients. Therefore, they may face more barriers to accessing the health system than other SUD patients. Moreover, heroin users tend to use mostly emergency care services, for blood-borne/local infection and traumatic injury [24].

International research on people with SUD and COVID-19 infection has shown a greater risk of infection and worse outcomes for people with SUD infected with COVID-19. A retrospective study carried out in the USA, which compared the incidence risk for COVID-19 in the general population with that for people with SUD, showed that people with a recent diagnosis of SUD had up to nine times the probability of infection than the general population; and that those who entered treatment for opioid abuse were the group at highest risk of infection [13]. However, it should be noted that the situation and profile of opioid users in Catalonia is very different to in the USA where there is a synthetic opioids crisis [25], and in the current study in Catalonia patients with opioid use disorder only included heroin users. In Catalonia, those seeking treatment for heroin have a worse socioeconomic profile than those seeking treatment for other substances, with 32.98% having a very low socioeconomic level, and 42.70% not living in stable housing. Often this socioeconomic disadvantage is coupled with greater social isolation [26], a factor which could have acted as a protective factor for COVID-19 infection in Catalonia among heroin patients.

Similar results to the current study were seen in a Brazilian study, which showed that people with alcohol use disorder had greater risk of infection than those with cocaine/crack use disorder and other groups [27]. Numerous studies demonstrate the impact of heavy alcohol use on the immune system, increasing vulnerability to respiratory infections [28]. This may contribute to the greater risk of COVID-19 infection among those seeking treatment for alcohol.

Other studies have shown that women, people with a low SEP and migrants are at greater risk of infection [29‒31]. Nonetheless, in the current study of people with SUD, an association was only seen between being a migrant and COVID-19 infection. This may be explained by them having more comorbidities: diabetes, cardiac diseases, asthma, HIV, and obesity; and by this population having more precarious employment and economic vulnerability, coupled with worse living conditions and overcrowding making it difficult to maintain physical distance and confinement [29‒32]. The lack of association between SEP and COVID-19 infection found in the present study could be explained because SUD patients’ socioeconomic level is lower than the socioeconomic level of general population (21.21% of SUD patients have very low SEP and 63.70% have a low SEP).

During the pandemic, people without stable housing faced more difficulties following the protection measures imposed by the government and were therefore at greater risk of exposure to COVID-19. In addition, these people are more clinically vulnerable and have more physical and mental comorbidities and high rates of substance abuse [33‒35]. Nonetheless, in the current study, as in other studies of COVID-19 incidence among homeless people, a lower risk of infection was observed in this group compared to those living in residential institutions. One study in Denmark showed that very few homeless people had COVID-19 antibodies [35], suggesting that social isolation could have been a protective factor in this group. The results of this current study, which observed higher risk among people in residential institutions, are similar to other studies [36‒38]. Prisons and other institutional centres are epicentres for infection, as they have high numbers of people with higher susceptibility and more difficulty in maintaining physical distance [38, 39].

Among the analysed clinical variables, this study showed an association between HIV and COVID-19 infection, with people with both SUD and HIV showing a greater risk of infection. A study in Catalonia which analysed COVID-19 infection in people with HIV found an incidence of 5.7% between March and December 2020, higher than for the general population. Being a migrant, being a man who has sex with men, or having four or more comorbidities (renal, respiratory, metabolic, autoimmune or cardiovascular), were associated with higher risk of COVID-19 infection [40]. Therefore, this study shows that a higher AMG rank is associated with higher cumulative incidence of COVID-19.

Although people with SUD have many risks associated with COVID-19 infection and poorer prognosis (higher rates of hospitalization, use of ventilation and death), compared to general population [12, 14, 41], the current study does not show higher cumulative incidence of infection. An observational study done in the Hospital del Mar, a reference hospital in an area of Barcelona with a high prevalence of people with SUD, showed that very few people with SUD were admitted to the hospital during the study period [42].

Considering this, to understand the impact and consequences of COVID-19 on people with SUD in Catalonia it is important to look further into COVID-19 and people with SUD. This knowledge will allow designing public health strategies which ensure equity and contribute to preventing avoidable health inequalities, prioritising screening, access to the health system and care for those most at risk of infection and of suffering clinical complications.

This is a patient-based study with many strengths. Catalonia has a universal public health system with valid and robust health registries. These contain sociodemographic, clinical and substance use-related information regarding people who seek treatment for SUD and who represent more than 3% of the population. The CHSS includes results of all diagnostic tests done of suspected cases of COVID-19. To date, we do not know of other studies at national or European level which have analysed the incidence of COVID-19 among people with SUD and the associated factors for infection in this group. Moreover, analysing the cumulative incidence according to substance may help generalize the results obtained to another SUD patient populations. Nonetheless, further studies are needed to understand the differences found between people with SUD.

Limitations that should be mentioned include that the pandemic saturated the Catalan health system, which was critical in the first quarter, and this resulted in under-detection of suspected and positive cases. A further limitation is the impossibility of following up patients in SUD treatment without a health identifier. The particular situation of these individuals and their sociodemographic characteristics may have contributed to a different response to COVID-19 infection. Also, some people sought treatment in private centres and these were not included in this study.

The study used de-identified, retrospective data, and was conducted according to the World Medical Association Declaration of Helsinki (ethical principles for medical research involving human subjects https://www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles-for-medicalresearch-involving-human-subjects/). The study protocol was approved by the Ethics Committee of the Germans Trias i Pujol Hospital (CEI-PI-20-221, Badalona, Spain), which waived the need for written informed consent.

The authors have no conflicts of interest to declare.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Irene Lana-Lander, Regina Muñoz-Galán, and Jorge Palacio-Vieira reviewed the literature and conceived and designed the study. Regina Muñoz-Galán and Elisenda Martínez-Carbonell performed the data extraction. Irene Lana-Lander and Regina Muñoz-Galán cleaned and analysed the data and drafted the initial version of the manuscript. Irene Lana-Lander, Regina Muñoz-Galán, Jorge Palacio-Vieira, Robert Muga, Xavier Majo-Roca, and Joan Colom reviewed the initial draft, made critical contributions to the interpretation of the data and approved the manuscript. The corresponding author attests that all listed authors meet the authorship criteria and that no others meeting the criteria have been omitted.

The datasets made for this study can be available on demand to the corresponding author. Part of the data could not be allowed for distribution to others than the investigation group as it may violate ethical and/or legal regulations of written consent.