Prevalence and risk factors of PTSD symptoms: a 3-month follow-up study

Abstract

Background:

During the peak of the epidemic, hospitalized patients frequently encountered significant health risks and potentially life-threatening circumstances, including uncertainty regarding treatment and the potential for complications.

Objective:

The present study aimed to explore the prevalence of post-traumatic stress disorder (PTSD) symptoms among hospitalized patients 3 months after discharge during the first peak of the epidemic, and the association of PTSD with disease-related characteristics.

Design:

A single-center and full-sample follow-up study was conducted on COVID-19 patients from the Optical Valley Branch of Maternal and Child Hospital of Hubei Province, Wuhan, China. Data were collected during their hospitalization and 3 months after discharge.

Methods:

PTSD symptoms were evaluated by primary care post-traumatic stress disorder (PC-PTSD), a total score of 3 or above was considered as clinically significant PTSD symptoms. Demographic and disease-related characteristics were collected to identify related associations with PTSD symptoms.

Results:

A total of 903 patients completed the follow-up survey, yielding a response rate of 63.5%. A total of 212 (23.5%) of the patients were positive in PC-PTSD screening. Univariate regression analysis identified several factors correlated with PTSD symptoms, including female gender, younger age, a lower body mass index (BMI), preexisting sleep problems, bereavement due to COVID-19, a severe clinical diagnosis, the presence of three or more clinical symptoms at disease onset, and residual respiratory symptoms after discharge. Notably, in the multivariate regression analysis, experiencing three or more clinical symptoms at onset emerged as a robust predictor of PTSD symptoms (OR = 2.09, 95% CI: 1.48–2.95). An intriguing finding was that patients who underwent radiological assessment post-discharge reported a higher incidence of PTSD symptoms, whereas those who underwent re-testing for IgG or IgM antibodies exhibited a lower prevalence of PTSD symptoms.

Conclusion:

Three months post-recovery, PTSD symptoms prevalence among COVID-19 patients was 23.5%. Those with three or more clinical symptoms at onset or residual respiratory symptoms post-discharge showed higher risk. These findings highlighted the long-term effect of COVID-19 on mental health, urging enhanced attention and interventions for survivors.

Plain language summary

Prevalence and risk factors of PTSD symptoms: a three-month follow-up study

Why was the study done?: PTSD (Post-Traumatic Stress Disorder) is a mental health condition that can develop after someone goes through or witnesses an extremely frightening, life-threatening event. It’s a type of ongoing mental distress that can show up later and last for a long time, significantly affecting a person’s quality of life. The study aimed to find out how many people who have gone through a public health emergency like COVID-19 develop PTSD symptoms within three months of recovery, and see if there’s a link between these symptoms and their illness. What did the researchers do? The research team did a single-center and full-sample follow-up study in COVID-19 patients who had recovered for three months. A tool called the PC-PTSD scale was used to check if they had PTSD symptoms. If their score was 3 or above, it meant they might have significant PTSD symptoms. The details about their illness were also collected to see if there’s a connection with PTSD symptoms. What did the researchers find? A total of 903 patients participated in the survey, with a participation rate of 63.5%. Among them, 212 patients (23.5%) had a positive result on the PC-PTSD test, indicating they might have PTSD symptoms. The study found that several factors were linked to PTSD symptoms, such as being female, younger age, having a low body weight, previous sleep issues, losing a loved one due to COVID-19, having a severe COVID-19 diagnosis, experiencing three or more symptoms at onset, and having residual respiratory issues after discharge. Specifically, those who had three or more symptoms at onset were more likely to develop PTSD. What do the findings mean? Within three months of COVID-19 recovery, 23.5% of patients may experience PTSD symptoms. Those who had three or more symptoms at onset or had residual respiratory issues after discharge were more likely to develop PTSD symptoms. This suggests that COVID-19 can have long-term psychological effects on patients.

Introduction

The COVID-19 outbreak appeared at the end of 2019 and soon became a public health emergency of international concern. ¹ Such pandemics with high infectivity and mortality have already proved to generate psychological impact, ² including post-traumatic stress disorder (PTSD). PTSD is a long-term psychiatric condition caused by traumatic occurrences such as natural disasters, traffic accidents, and violence, both direct and indirect. ³ Its typical symptoms are intrusions, avoidance, negative mood or thoughts, and hyperarousal. This emotional suffering could last 10 years or longer, ⁴ severely affecting one’s life quality, ⁵ and even changing individual characteristics. ⁶ The psychological effects of such pandemics were delayed even after recovery. For example, some developed persistent psychological distress and diminished social functioning at 4 years post-SARS. ⁷ Other infectious diseases like MERS ⁸ also have a similar tendency.

Previous studies showed that both patients of COVID-19 and residents of Wuhan developed PTSD symptoms (measured by the PTSD Checklist for DSM-5) like sleeping difficulty^9,10 during the initial surge of the COVID-19 pandemic. Our previous research found that nearly a quarter of COVID-19 patients reported high levels of PTSD symptoms during hospitalization, ¹¹ and a recent review showed that the high incidence of PTSD symptoms persisted for 1 year after discharge. ¹² In a cohort follow-up study of 115 recovered COVID-19 patients in Italy, ¹³ 10.4% met a PTSD Checklist for DSM-5 (PCL-5)-based diagnosis of PTSD and 8.6% had subthreshold PTSD at 3 months. In an observational study of 239 patients diagnosed with COVID-19 in the Netherlands, the results showed that 37.2% of patients had PTSD symptoms at 3 months. ¹⁴ A 3-month follow-up study of 318 COVID-19 survivors in Peru found that 29.5% of patients had PTSD symptoms. An increasing number of researchers come to the same conclusion that COVID-19 patients are at a high risk of PTSD, ¹⁵ and the risk persisted for a longer period of time. Three months seemed to be an important time point for developing PTSD under COVID-19.

Although, with PTSD, one would experience chronic psychological pain, it still takes a long time for one to seek help for mental health problems. As mentioned in previous research, the median span for those who seeking psychological support was 36 months, while those who with no experience may need a longer time, approximately 64 months. ¹⁶ This largely reduced one’s life quality and affected the prognosis of treatments. Therefore, it was necessary for healthcare institutions to conduct screening for PTSD symptoms at different time points after discharge for COVID-19 hospitalized patients so that timely psychological intervention and assistance could be provided to those in need. In addition, exploring the factors influencing PTSD symptoms could help us find people at risk and provide them with the right help. Previous researchers found that the prevalence of PTSD for women^17,18 was nearly twice as high as for men after surviving a natural disaster and similar conclusions had been drawn in a SARS-related research. ¹⁷ Also, data from several studies suggested ⁹ that women had more significant PTSD symptoms (PTSS), for instance, negative alterations in cognition or mood and re-experiencing. Besides, certain somatic symptoms such as cardio-respiratory, gastrointestinal health, and musculoskeletal pain ¹⁹ were claimed to be associated with this terrible disorder. Up to now, many reports indicated that medical staff, residents in mandatory quarantine zones, ²⁰ patients who have had PTSS before, and females as well as those with sleep problems were more vulnerable during the pandemic. ⁹ Looking for relevant influencing factors would be helpful for future timely psychological intervention and assistance.

PTSD was considered the second tsunami of COVID-19. ²¹ The COVID-19 pandemic was not only a public health crisis but also a mental health crisis. PTSD, as one of the common psychological disorders after the pandemic, deserved global attention and recognition due to its high prevalence and widespread impact. Therefore, there was an urgent need to understand the prevalence and risk factors of PTSD associated with persistent symptoms after COVID-19, to assist healthcare institutions in developing better strategies to manage these patients. However, most studies on PTSD symptoms in COVID-19 patients were cross-sectional, lacking longitudinal research on the development of PTSD symptoms during the recovery period. In the present study, we conducted a large-sample longitudinal study to explore the PTSD symptoms 3 months after the discharge and its relationship with disease-related characteristics. Related results of our study can provide a reference for early identification of COVID-19 patients at risk of PTSD symptoms so that necessary and timely help can be provided.

Methods

The reporting of this study conformed to the Strengthening the Reporting of Observational Studies in Epidemiology statement. ²²

Study design and participants

This follow-up study was designed to track the prevalence and possible risk factors of PTSD symptoms in COVID-19 patients 3 months after discharge. The inclusion criteria were as follows: (1) age ⩾ 18 years old; (2) laboratory-confirmed with COVID-19; (3) clear and complete medical information; (4) complete contact information; and (5) no serious mental illness before. The exclusion criteria were as follows: (1) incorrect or unreachable contact information; (2) unwillingness to participate in the study; and (3) an uncertain history of mental illness, characterized by gaps or inconsistencies in medical records regarding previous diagnoses or treatments. Data were extracted from electronic medical records managed by the Optical Valley Branch of Maternal and Child Hospital of Hubei Province. Between February 20 and March 31 2020, 1767 suspected COVID-19 patients were admitted to the hospital, and 277 patients were negative in results on nucleic acid and specific serum antibody detection during hospitalization. A total of 58 patients lacked contact information, 7 patients lacked medical data, 5 patients were less than 18 years old, and a total of 1420 COVID-19 patients were included in the follow-up list. The research team designed a standardized electronic follow-up questionnaire for the follow-up 3 months after discharge. All patients were followed up by 15 clinicians through phone calls after discharge. Those who did not respond to the initial call were contacted again at different times to minimize the dropouts. In the follow-up survey, 281 patients were lost to follow-up due to refusing to participate or wrong contact information and 212 patients did not respond before the deadline. A total of 9 patients were excluded due to uncertain history of mental illness, and 15 patients lacked results of PC-PTSD. Finally, a total of 903 patients entered the statistical analysis, with a response rate of 63.5%.

Measures

A standardized questionnaire containing demographic and clinical information as well as PTSD screening was designed to ensure accurate and comprehensive information collection. A copy of the questionnaire was provided in the supplementary material (Appendix 1 Questionnaire).

Demographic and disease-related characteristics

Demographic variables were extracted on gender (female and male), age (18–44, 45–64, ⩾65 years), and BMI (<18.5, 18.5–23.9, >23.9 kg/m²). History of tobacco and alcohol use and sleep problems as previous personal history were also collected. In the clinical-related characteristics, we speculated that more clinical symptoms and more severe clinical diagnosis may mean a higher degree of trauma exposure, which would lead to more PTSD symptoms. So, we collected the severity of COVID-19 diagnosed during hospitalization and whether there were clinical symptoms at the time of onset. For information after discharge, we hypothesized that the remaining symptoms and medical tests after discharge may be reminders of previous trauma, so we collected the residual respiratory symptoms such as chronic cough and dyspnea and medical reexamination.

PTSD symptoms screening

PC-PTSD (primary care PTSD screen) was used to identify PTSD symptoms, which was developed to quickly detect PTSD based on DSM-IV PTSD diagnostic criteria. ²³ The PC-PTSD has been used widely in primary care and mental health care settings due to its high diagnostic validity.^24–26 In comparison with the CAPS (Clinician-Administered PTSD Scale) diagnosis, the sensitivity of the PC-PTSD was 0.78 and the specificity was 0.87. The PC-PTSD included four items, and each item was designed to report whether the following symptoms were present or not, including reexperiencing, avoidance, hyperarousal, and numbing. Answering “yes” was scored as 1, answering “no” as 0, and the scores of four items were added to get a total score. Generally, a total score of 3 or above is considered a positive result, indicating clinically significant PTSD symptoms. The Chinese version of PC-PTSD has good reliability and validity among medical staff and the general population.^27,28 The Cronbach’α of PC-PTSD in this study was 0.914.

Statistical analysis

For descriptive analyses, frequencies were calculated. We used the chi-squared test to assess differences in the prevalence of PTSD and the four main PTSD symptoms across demographic variables and disease-related characteristics, with odds ratio and 95% confidence interval calculated. A multivariable-adjusted logistic regression analysis was used to explore the risk factors associated with the prevalence of PTSD and its four cardinal symptoms. Four primary models were used to examine the effect of multivariable on PTSD, with adjustment for varying levels of covariates. In the first model, demographic and lifestyle variables, including gender, age, body mass index (BMI), smoking and drinking history, as well as preexisting sleep problems were incorporated into the regression equation to establish a foundational understanding of their relationships with PTSD. Subsequently, the second model expanded upon this foundation by incorporating the number of clinical symptoms at onset and the severity of COVID-19 infection to ascertain whether, after controlling for demographic variables, the severity of the illness itself had a statistically significant impact on the subsequent occurrence of PTSD. In the third model, we included the variables after discharge, including residual respiratory symptoms and reexamination, into the equation to investigate the potential influence of these post-discharge factors on the development of PTSD. Finally, the loss of a loved one because of COVID-19 was included in the regression analysis to see the cumulative effect of all variables. Statistical analyses were performed in SPSS, version 22.0 (IBM SPSS Inc., Chicago, USA). A two-sided p < 0.05 was considered significant.

Results

Sample description

A total of 60.0% of the 903 patients were female, and their median age was 58.5 years old (IQR 49–67, see Table 1). A total of 35 individuals experienced the loss of a loved one due to COVID-19, while 80 patients reported preexisting sleep disturbances, which were more prevalent among females (see Supplemental eTable 1). Clinical severity was diagnosed in 43 patients (4.8%), with a significantly higher proportion among males compared to females (7.8% vs 2.8%; chi-square = 11.732, p < 0.001). At disease onset, 88.2% of patients manifested at least one symptom, with fever, cough, and fatigue being the most frequently reported. Notably, women reported a higher incidence of fatigue symptoms than men (40.6% vs 28.8%; chi-square = 13.073, p < 0.001). Post-discharge, 54 patients (6.0%) continued to experience chronic cough, while 64 patients (7.1%) suffered from varying degrees of dyspnea. Furthermore, 19 patients (2.1%) tested positive for nucleic acids post-discharge, and 61.4% of the cohort underwent retesting for IgG or IgM antibodies. 461 (51.1%) patients underwent a post-discharge radiologic assessment and 44 (4.9%) patients underwent a pulmonary function test.

Table 1.

Demographic and disease-related characteristics of 903 patients with COVID-19.

Variables	N	%
Gender
Male	361	40.0
Female	542	60.0
Age, years
18–44	166	18.4
45–64	428	47.4
⩾65	309	34.2
BMI, kg/m2
⩽18.5	29	3.2
18.5–23.9	462	51.2
>23.9	409	45.3
Missing	3	0.3
Lost a loved one because of COVID-19	35	3.9
Having sleep problems before	80	8.9
Smoking history	120	13.3
Drinking history	71	7.9
Severity of COVID-19
Non-severe	854	94.6
Severe	43	4.8
Missing	6	0.7
Symptoms at onset
Any	796	88.2
Fever	537	59.5
Cough	458	50.7
Dyspnea	170	18.8
Fatigue	324	35.9
Gastrointestinal symptoms	111	12.3
Anorexia	191	21.2
Myalgia or arthralgia	138	15.3
Headache	36	4.0
Loss of taste or dysgeusia	57	6.3
Loss of smell or dysosmia	27	3.0
Number of symptoms at onset
<3	559	61.9
⩾3	344	38.1
Symptoms after discharge
Cough	54	6.0
Dyspnea	64	7.1
Reexaminations after discharge
Positive in nucleic acid test	19	2.1
Reexamine IgG and IgM	554	61.4
Taking radiological examination	461	51.1
Retesting of lung function	44	4.9
PC-PTSD items
Reexperiencing	265	29.3
Avoidance	230	25.5
Hyperarousal	245	27.1
Numbing	250	27.7
PC-PTSD scores
0	561	62.1
1	70	7.8
2	60	6.6
3	48	5.3
4	164	18.2
Positive in PC-PTSD result	212	23.5

BMI, body mass index; PC-PTSD, primary care PTSD screen; PTSD, post-traumatic stress disorder.

Associations between demographic and disease-related characteristics and PC-PTSD results

The PC-PTSD screening revealed that 212 patients (23.5%) exhibited clinically significant PTSD symptoms, with a higher prevalence among females compared to males (OR = 1.48, 95% CI 1.07–2.04; see Table 2). Notably, younger patients aged 18–44 years demonstrated a significantly elevated prevalence of PTSD symptoms (OR = 1.64, 95% CI 1.05–2.55) in contrast to older individuals over 65 years. In addition, patients with lower BMI, preexisting sleep disturbances (OR = 1.87, 95% CI 1.15–3.05), and those who experienced the loss of a loved one due to COVID-19 (OR = 4.15, 95% CI 2.10–8.23) exhibited a higher prevalence of significant PTSD symptoms. Clinical severity of COVID-19 diagnosis also played a role, with patients diagnosed as severe exhibiting a higher prevalence of PTSD compared to those with non-severe diagnoses (OR = 2.02, 95% CI 1.06–3.82). The initial clinical symptoms such as dyspnea, fatigue, gastrointestinal disturbances, anorexia, headache, and alterations in taste or smell were associated with a higher prevalence of PTSD (see Figure 1). Post-discharge, patients who continued to experience chronic cough (OR = 2.01, 95% CI 1.13–3.58) and dyspnea (OR = 1.79, 95% CI 1.04–3.07) had a significantly higher prevalence of PTSD symptoms compared to those who did not report these symptoms. Interestingly, patients who underwent retesting for IgG or IgM antibodies post-discharge exhibited a lower prevalence of PTSD symptoms, whereas those who underwent radiological examinations or lung function tests demonstrated a higher prevalence of significant PTSD symptoms.

Table 2.

Demographic characteristics and disease-related characteristics associated with positive PC-PTSD results.

Variables	PTSD negative		PTSD positive		χ2	p Value	OR	95% CI
	N	%	N	%
Gender
Male	291	80.6	70	19.4	5.591	0.018*	Ref	—	—
Female	400	73.8	142	26.2			1.48	1.07	2.04
Age, years
18–44	119	72.1	47	27.9	5.490	0.064	1.64	1.05	2.55
45–64	322	75.2	106	24.8			1.40	0.97	2.00
⩾65	250	80.9	59	19.1			Ref	—	—
BMI, kg/m2
<18.5	17	58.6	12	41.4	12.611	0.002**	3.04	1.40	6.64
18.5–23.9	339	73.4	123	26.6			1.56	1.13	2.16
>23.9	332	81.2	77	18.8			Ref	—	—
Lost a loved one because of COVID-19
No	675	77.8	193	22.2	19.237	<0.001***	Ref	—	—
Yes	16	45.7	19	54.3			4.15	2.10	8.23
Having sleep problems before
No	639	77.6	184	22.4	6.487	0.011*	Ref	—	—
Yes	52	65.0	28	35.0			1.87	1.15	3.05
Smoking history
No	596	76.1	187	23.9	0.538	0.463	Ref	—	—
Yes	95	79.2	25	20.8			0.84	0.52	1.34
Drinking history
No	633	76.1	199	23.9	1.145	0.285	Ref	—	—
Yes	58	81.7	13	18.3			0.71	0.38	1.33
Severity of COVID-19
Non-severe	660	77.3	194	22.7	4.796	0.029*	Ref	—	—
Severe	27	62.8	16	37.2			2.02	1.06	3.82
Symptoms at onset
Any
No	87	81.3	20	18.7	1.547	0.214	Ref	—	—
Yes	604	75.9	192	24.1			1.38	0.83	2.31
Fever
No	284	77.6	82	22.4	0.394	0.530	Ref	—	—
Yes	407	75.8	130	24.2			1.11	0.81	1.52
Cough
No	350	78.7	95	21.3	2.214	0.137	Ref	—	—
Yes	341	74.5	117	25.5			1.26	0.93	1.72
Dyspnea
No	575	78.4	158	21.6	8.006	0.005**	Ref	—	—
Yes	116	68.2	54	31.8			1.69	1.17	2.45
Fatigue
No	466	80.5	113	19.5	14.092	<0.001***	Ref	—	—
Yes	225	69.4	99	30.6			1.81	1.33	2.48
Gastrointestinal symptoms
No	626	79.0	166	21.0	22.733	<0.001***	Ref	—	—
Yes	65	58.6	46	41.4			2.67	1.76	4.04
Anorexia
No	570	80.1	142	19.9	23.394	<0.001***	Ref	—	—
Yes	121	63.4	70	36.6			2.32	1.64	3.29
Myalgia or arthralgia
No	549	71.8	216	28.2	2.982	0.084	Ref	—	—
Yes	89	64.5	49	35.5			1.40	0.95	2.05
Headache
No	670	77.3	197	22.7	6.905	0.009**	Ref	—	—
Yes	21	58.3	15	41.7			2.43	1.23	4.80
Loss of taste or dysgeusia
No	657	77.7	189	22.3	9.642	0.002**	Ref	—	—
Yes	34	59.6	23	40.4			2.35	1.35	4.09
Loss of smell or dysosmia
No	675	77.1	201	22.9	4.617	0.032*	Ref	—	—
Yes	16	59.3	11	40.7			2.31	1.05	5.05
Number of symptoms at onset
<3	465	83.2	94	16.8	36.245	<0.001***	Ref	—	—
⩾3	226	65.7	118	34.3			2.58	1.89	3.54
Symptoms after discharge
Cough
No	657	77.4	192	22.6	5.878	0.015*	Ref	—	—
Yes	34	63.0	20	37.0			2.01	1.13	3.58
Dyspnea
No	649	77.4	190	22.6	4.553	0.033*	Ref	—	—
Yes	42	65.6	22	34.4			1.79	1.04	3.07
Reexaminations after discharge
Positive in nucleic acid test
No	680	76.9	204	23.1	2.764	0.096	Ref	—	—
Yes	11	57.9	8	42.1			2.42	0.96	6.11
Reexamine IgG and IgM
No	253	72.5	96	27.5	5.142	0.023*	Ref	—	—
Yes	438	79.1	116	20.9			0.70	0.51	0.95
Taking radiological examination
No	369	83.5	73	16.5	23.355	<0.001***	Ref	—	—
Yes	322	69.8	139	30.2			2.18	1.58	3.01
Retesting of lung function
No	666	77.5	193	22.5	9.996	0.002**	Ref	—	—
Yes	25	56.8	19	43.2			2.62	1.41	4.86

The parameters marked in bold represent statistical significance. *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001.

BMI, body mass index; PC-PTSD, primary care PTSD; PTSD, post-traumatic stress disorder.

Figure 1.

The odds ratio of clinical symptoms at the onset of COVID-19 and after discharge on the incidence of PTSD symptoms: (a) The prevalence of PTSD symptoms was higher among patients having clinical symptoms at the onset of COVID-19: dyspnea, fatigue, gastrointestinal symptoms, anorexia, headache, and loss of taste or loss of smell and (b) patients with chronic cough or dyspnea after discharge, and who retesting radiological examination or lung function had a higher prevalence of PTSD symptoms. While patients who retested IgG or IgM after discharge had a lower prevalence.

PTSD, post-traumatic stress disorder.

Associations between demographic and disease-related characteristics and four PTSD symptoms

The PC-PTSD was employed to examine four primary symptoms of PTSD: reexperiencing, avoidance, hyperarousal, and numbing. Among the study cohort, a substantial proportion of patients reported experiencing these symptoms, with 265 (29.3%) endorsing reexperiencing, 230 (25.5%) avoidance, 245 (27.1%) hyperarousal, and 250 (27.7%) numbing. It is noteworthy that the prevalence of reexperiencing, avoidance, and hyperarousal symptoms was significantly elevated in female patients compared to their male counterparts, whereas no such gender disparity was observed for numbing symptoms. All four PTSD symptoms were associated with initial clinical symptoms of COVID-19, such as fatigue, gastrointestinal symptoms, or anorexia, and were more common in patients with three or more initial clinical symptoms (see Supplemental eTable 2 and eFigure 1). Moreover, our analysis demonstrated that patients who reported a chronic cough following their discharge were more susceptible to reexperiencing, avoidance, and numbing symptoms(see Figure 2 and Supplemental eFigure 2). Patients who tested positive for nucleic acid after discharge were more likely to experience re-experiencing, avoidance, and hyperarousal symptoms. On the other hand, those who retested IgG or IgM after discharge had fewer re-experiencing, avoidance, and numbing symptoms. Patients who retested radiological examination or lung function after discharge developed more avoidance, hyperarousal, and numbing symptoms (see Supplemental eTable 2).

Figure 2.

Associations between demographic and disease-related characteristics and four PTSD symptoms: (a) Four PTSD symptoms investigated by PC-PTSD were reexperiencing, avoidance, hyperarousal, and numbing. They were all associated with clinical symptoms at the onset of COVID-19 such as fatigue, gastrointestinal symptoms, anorexia, or with three or more clinical symptoms and (b) the prevalence of reexperiencing, avoidance, and hyperarousal symptoms was higher for female patients than males, but not in numbing symptoms.

PTSD, post-traumatic stress disorder; PC-PTSD, primary care PTSD.

Regression models

Table 3 presents a comprehensive overview of the findings from a series of multivariable-adjusted logistic regression analyses conducted to investigate the predictors of PTSD symptoms among the studied population. In Model 1, younger age (below 65 years), lower body mass index (BMI, <23.9 kg/m²), and pre-existing sleep disturbances were identified as significant factors associated with an increased likelihood of PTSD symptoms. No notable differences were observed for gender, smoking, or alcohol consumption histories. In model 2, patients with a clinical diagnosis of severe (OR = 2.16, 95% CI 1.07–4.34) or with more than three symptoms at the onset of COVID-19 (OR = 2.38, 95% CI 1.72–3.31) were more likely to suffer from PTSD symptoms after controlling for all covariates in model 1. In Model 3, which further controlled for demographic, sociological factors, and disease severity variables, undergoing a radiological examination post-discharge was positively associated with PTSD (OR = 2.01, 95% CI 1.41–2.85). Conversely, patients who underwent post-discharge retesting for IgG or IgM antibodies demonstrated a lower prevalence of PTSD (OR = 0.66, 95% CI 0.47–0.93). In the most comprehensive Model 4, which incorporated all covariates, younger age, lower BMI, experiencing more than three symptoms at the onset of the illness, radiological re-examination, IgG/IgM retesting, and the loss of a loved one due to the illness were found to be significantly correlated with the development of PTSD. However, when Model 4 was applied to male patients specifically (see Supplemental eTable 3 and 4), the significance of certain variables, including symptom count, radiological and serological retesting, and bereavement due to COVID-19, was diminished.

Table 3.

Regression models of demographic and disease-related characteristics on PC-PTSD positive result.

	Model 1			Model 2			Model 3			Model 4
	OR	95% CI		OR	95% CI		OR	95% CI		OR	95% CI
		Lower	Upper		Lower	Upper		Lower	Upper		Lower	Upper
Gender
Female vs Male	1.40	0.94	2.07	1.46	0.97	2.18	1.49	0.98	2.25	1.45	0.96	2.21
Age, years
18–44 vs ⩾65	1.75*	1.11	2.77	1.83*	1.14	2.94	2.05**	1.25	3.34	2.14**	1.30	3.50
45–64 vs ⩾65	1.47*	1.02	2.14	1.49*	1.02	2.20	1.52*	1.02	2.25	1.58*	1.06	2.36
BMI, kg/m2
<18.5 vs >23.9	3.51**	1.57	7.84	3.39**	1.47	7.80	3.39**	1.43	8.00	3.67**	1.54	8.77
18.5–23.9 vs >23.9	1.54*	1.10	2.15	1.63**	1.16	2.29	1.69**	1.19	2.39	1.75**	1.23	2.49
Smoking history	1.23	0.68	2.22	1.42	0.77	2.59	1.36	0.74	2.50	1.34	0.73	2.49
Drinking history	0.86	0.41	1.77	0.78	0.37	1.64	0.76	0.36	1.62	0.79	0.37	1.69
Having sleep problems before	1.86*	1.13	3.07	1.65	0.98	2.76	1.51	0.88	2.56	1.45	0.85	2.49
Severity of COVID-19
Severe vs non-severe				2.16*	1.07	4.34	1.80	0.86	3.76	1.81	0.86	3.80
Number of symptoms at onset
⩾3 vs 0–2				2.38***	1.72	3.31	2.07***	1.47	2.91	2.09***	1.48	2.95
Symptoms after discharge
Cough							1.24	0.64	2.40	1.30	0.67	2.51
Dyspnea							1.14	0.61	2.14	1.15	0.61	2.17
Reexaminations after discharge
Positive in nucleic acid test							1.77	0.65	4.82	1.64	0.60	4.51
Reexamine IgG and IgM							0.66*	0.47	0.93	0.67*	0.48	0.94
Taking radiological examination							2.01***	1.41	2.85	1.98***	1.39	2.82
Retesting of lung function							1.58	0.78	3.17	1.50	0.73	3.06
Lost a loved one because of COVID-19										4.13***	1.99	8.59

The parameters marked in bold represent statistical significance. * p-value < 0.05; **p-value < 0.01; ***p-value < 0.001.

BMI, body mass index; PC-PTSD, primary care PTSD screen.

In addition, separate regression analyses targeting the four cardinal PTSD symptoms (reexperiencing, avoidance, hyperarousal, and numbing; detailed in Supplemental eTable 5–eTable 8) revealed that younger age, lower BMI, multiple symptoms at COVID-19 onset, and loss of a loved one were consistently associated with all four symptoms. Notably, radiological and serological retesting post-discharge were significantly linked to reexperiencing and avoidance symptoms, but not to hyperarousal or numbing symptoms, when all covariates were considered.

Discussion

The current study explored the PTSD symptoms 3 months after the discharge of COVID-19 patients and its relationship with disease-related characteristics. Results indicated that 23.5% of patients experienced probable PTSD symptoms according to PC-PTSD. Female, younger age, lower BMI, previous sleep problems, loss of relatives due to COVID-19, more clinical symptoms at the onset of COVID, severe diagnosis, and respiratory symptoms after discharge were all risk factors for PTSD symptoms.

Numerous studies have been conducted to investigate the physical and psychological conditions of patients with COVID-19. Ju et al. ²⁹ reported that the prevalence of PTSD was 36% in a follow-up study of COVID-19 patients 1 month after discharge. In an Italian cohort study, the prevalence of PTSD in COVID-19 patients at 4 months after discharge from the hospital was 17.2%. ³⁰ These findings were in line with the possible occurrence of PTSD which was reported in the current study. The PC-PTSD scale was more concise than other PTSD screening instruments, such as the PCL-5. Furthermore, the sensitivity and specificity of the PC-PTSD scale varied when different cutoff values were employed for screening. A previous study demonstrated that the heterogeneity of measurement tools contributed to the observed differences in the prevalence rate of PTSD post-COVID-19. ³¹ A study among patients with substance use disorder revealed that the prevalence rate of PTSD was 44.4% when the PCL-C scale was employed with a cutoff score of 30, and 49.7% when the PC-PTSD scale was utilized with a cut-off score of 3. ²⁴ In other studies validating the PC-PTSD in different populations, no statistical differences in the detection rates of the scale were reported in comparison to other PTSD screening scales, ³² with the Area Under the ROC Curve (AUC) ranging from 0.8 to 0.95. ²³

The severity of the disease was positively related to the prevalence of PTSD in the present study. First, severe initial symptoms may trigger a profound stress response in the body, including the release of stress hormones like cortisol and adrenaline. This hyperactivation of the stress response system could lead to long-term alterations in brain chemistry and function, particularly in areas associated with emotion regulation and memory processing, which were key components of PTSD. Second, due to the nature of critical illness, patients were often separated from their loved ones and familiar surroundings. This isolation may lead to feelings of loneliness, abandonment, and fear, which could further exacerbate PTSD symptoms. The lack of family support ³³ (which cannot be replaced by the support of friends) and the increased loneliness, coupled with the guilt of not knowing whether family members have been infected by themselves, were associated with the prevalence of PTSD.

The more physical symptoms, the more psychological effects. The current study pointed out that among COVID-19 patients, those who initially presented with three or more clinical symptoms were more likely to develop PTSD symptoms after recovery. This indicated that early identification of high-risk groups and timely psychological intervention were crucial. Healthcare professionals should assess patients’ psychological status upon admission and provide necessary psychological support and intervention measures for those at high risk. Patients with fatigue, gastrointestinal symptoms, anorexia, and loss of taste or dysgeusia at onset were more likely to have all four symptoms of PTSD (Figure 1 and Supplemental eFigure 2). Meanwhile, patients with chronic cough and dyspnea after discharge were more vulnerable to avoidance and numbing symptoms (Supplemental eFigure 1). The result showed that chronic symptoms at onset or after discharge may have a deeper impact on patients, bringing them back to the terrible memory of that difficult time. Relevant institutions should provide certain psychological dredge and social support to patients according to their symptoms at onset and after discharge. Timely and effective psychological interventions could improve the quality of life of high-risk groups. Surprisingly, patients who reexamined IgM and IgG after discharge had a lower PTSD prevalence. Repeat testing may reassure them that they were not re-infected. In addition, most people received the examination before returning to work, and it was believed that going back to work effectively would resume people’s social function and reduce the possibility of chronic mental disorders.

Although COVID-19 affected the same number of males and females, ³⁴ males showed higher mortality and vulnerability. But when it comes to the prevalence of PTSD, females seemed to be more sensitive. They would suffer from hyperarousal and reexperiencing, which certainly depressed their life satisfaction. Gender difference was confirmed again in line with what we have always known.^9,35 Age was also linked to the prevalence of PTSD. People between the ages of 18 and 44 (referred to as young adults) were most likely to suffer from PTSD. Studies found that high loneliness^36,37 and low distress tolerance^38,39 were associated with PTSS. Young adults were in the early stages of their studies or careers, which means they were relatively inexperienced in society. We speculated that being kept under mandatory quarantine or receiving closed treatments would aggravate their loneliness and make them more susceptible to PTSD. In addition, among all the patients, the underweight group (BMI < 18.5) was more likely to suffer from PTSD (Supplemental eFigure 3). This study adopted the Chinese BMI grading standard, so the BMI of the obese group we obtained (overweight > 23.9, obese > 28) was lower than the index used in Western studies (overweight > 25, obese > 30), so the result did not reveal the effect of obesity. In addition, there were certain differences between Eastern and Western cultures. In developing countries such as China, overweight was a symbol of health and wealth, and the overweight group had a greater sense of happiness. ⁴⁰ Therefore, in China, there was an expression “Xin Kuan Ti Pan.” After being infected with COVID-19, the poor cardiopulmonary function of underweight patients made it harder for them to overcome pulmonary infectious diseases. Malnutrition would lead to exercise limitations and functional impairment. ⁴¹ We fully agreed that overweight and obese individuals have a high prevalence of PTSD, which had been found by previous research.^42–44 However, we suggested that similar attention should be paid to underweight patients.

Limitations

Although this study found some major conclusions, it was important to acknowledge the limitations of the study. First, it should be noted that this was a single-center and full-sample study. Consequently, the findings may not be wholly representative of the experiences of patients with COVID-19 at other medical centers. Due to the disparate standards and patient populations admitted to different designated medical centers, there may be some sampling bias. Second, the prevalence of PTSD was analyzed using the PC-PTSD scale, with follow-up conducted via telephone rather than in person, in accordance with the prevailing social distancing measures. It should be noted that this was not a clinical assessment, and therefore the potential for subjective influences to affect the results cannot be discounted. Third, the lack of direct assessment of Criterion A in the PC-PTSD scale represented a significant limitation that must be acknowledged. It was hypothesized that during the initial stages of the pandemic, infected patients generally experienced severe health threats and potentially life-threatening situations, including prolonged isolation, uncertainty during treatment, and possible complications. The absence of direct assessment of Criterion A may result in the misclassification of individuals who have not experienced sufficiently severe trauma but exhibit similar symptoms as having PTSD. Therefore, when interpreting the results of our study, we must cautiously consider this limitation. Future studies may consider using a combination of questionnaires or clinical interviews that include an A-criterion assessment to more fully assess the development of PTSD symptoms.

Clinical implications

The present study suggested that about one in four COVID-19 patients suffered from PTSD 3 months after recovery. After the epidemic, the government or relevant departments should pay more attention to the COVID-19 patients’ mental health and take timely and effective intervention measures. However, it was more likely that these patients may not know how to seek professional help or avoid going to the hospital because of avoiding symptoms. Therefore, timely follow-up and identification of these patients to give timely help and guidance are particularly important.

In addition, our study attempted to explore the correlation between clinical symptoms and psychopathology, to help clinicians quickly identify the susceptible population of PTSD. Results showed that patients with gastrointestinal symptoms and fatigue at onset as well as patients diagnosed as severe were more likely to suffer from PTSD symptoms. So far, few previous studies have been conducted to explore the relationship between clinical symptoms and PTSD. To some extent, our study suggested a link between early clinical symptoms and the development of PTSD, and further research will be needed to confirm and explore the mechanism.