The coronavirus disease 2019 (COVID-19) pandemic has affected >770 million people worldwide, causing >7 million deaths as of 31 December 2023.1 The period between July and September 2020 constituted the third wave of the pandemic in Hong Kong, resulted in >1.2 million reported cases between 23 January 2020 and 29 January 2023.2 The containment strategies implemented during the third wave included mandatory mask wear in public places, even when exercising in public outdoor areas; suspensions of public leisure facilities and private gyms; and the initiation of work-from-home arrangements.3 These strategies led to reductions in physical activity and daily movement, with the goal of viral containment. Furthermore, compulsory social distancing and suspension of dine-in services were included among the policies that could affect various dietary and lifestyle habits, although these methods were less stringent than approaches in cities under lockdown. Overall, the unprecedented public health crisis created many challenges for Hong Kong residents attempting to maintain healthy lifestyle habits. Nevertheless, few studies have examined dietary and physical activity habits in the general population during the COVID-19 pandemic.4 5 6

Considering that individuals with chronic diseases are more likely to develop severe cases of COVID-19, this study aimed to measure the prevalences of unhealthy dietary habits and physical inactivity levels in an adult Chinese population, to identify factors associated with their adoption of these dietary and physical activity habits, and to perform a time trend analysis comparing the proportions of the population that adopted healthy dietary habits, physical activity levels, and avoidance of smoking and alcohol consumption during the third wave of the COVID-19 pandemic.

We utilised a methodology similar to a previous population-based, random telephone survey conducted in Hong Kong.7 Two-stage sampling was performed, in which participants were recruited by trained interviewers through a telephone interview system based on telephone calls to landlines identified by random digit dialling. The sample population was randomly selected by the Centre for Health Behaviours Research at The Chinese University of Hong Kong. Calls were made during typical office hours, 9 am to 5 pm, Monday through Saturday between 7 and 31 October 2020. Three attempts were made if the call initially was not answered. Territory-wide, any Chinese adults aged ≥18 years who could communicate in Chinese via telephone were eligible to participate. Assuming an outcome variable rate of 35%, at least 1456 participants were required to achieve a precision level of 2.5% from the following formula:
where ‘p’ stands for proportion and ‘N’ stands for sample size.

The interviews were performed using a fieldwork manual highlighting standard operating procedures by a team of trained interviewers and supervised by an experienced project coordinator throughout the study. The characteristics of survey participants are shown in Table 1.

The survey consisted of five sections: (1) socio-demographic details (age, sex, marital status, education level, job status, household income, and receipt of comprehensive social security assistance); (2) clinical information (eg, presence of chronic diseases); (3) smoking (current daily amount/ex-/non-smoker) and alcohol consumption habits (daily amount in the preceding 7 days); (4) dietary screening via the Hong Kong Diet Score (HKDS), using a validated scale that contained nine items assessing the participant’s daily consumption of nine food groups in the preceding 7 days; and (5) level of physical activity in the preceding 7 days, as determined by a Chinese version of the 7-item International Physical Activity Questionnaire Short Form (IPAQ-C).

The traditional Mediterranean diet is well-defined and has been positively associated with favourable health outcomes.8 9 10 The Mediterranean diet score is used to measure compliance with a traditional Mediterranean diet. This scoring system has been widely utilised in studies that measure Mediterranean diet adherence or adaptation as an indicator of healthy dietary choices. In this study, we developed the HKDS, a dietary screener that contained nine items assessing dietary intake of nine food groups (alcohol, legumes, grains, fruits, vegetables, meats, dairy, red/orange vegetables, and fatty fish) in the preceding 7 days. The screener incorporated key traditional Greek diet characteristics, known as the Mediterranean diet score of de Groot et al,8 which were also used in a study of Hong Kong Chinese by Woo et al (Table 2).11 The original 8-item survey was modified by removing the ratio of monounsaturated fatty acids to saturated fatty acids and replacing ethanol with alcohol. Dietary fatty acids and ethanol are widely distributed among various food groups; they are typically assessed through weighted foods, which are unlikely to be accurately determined using a single question in a telephone interview. Two additional items were included regarding carotenoid-rich and omega-3–rich food intake based on the Hong Kong Centre for Food Safety Recommended Nutrient Intake for vitamin A12 and the World Health Organization recommendation for omega-3. Both nutrients are inversely associated with incidence of non-communicable diseases (NCDs). For each item, consumption at or above the recommended amount was scored as 1 point and 0 points otherwise; however, for ethanol, consumption below the specified amount was scored as 1 point and 0 points otherwise. Each participant received a total score of 0 to 9; a score of ≥5 was considered high. A pilot survey was conducted with a convenience sample of 23 participants. Intraclass correlation coefficient estimates and 95% confidence intervals (CIs) were determined using a two-way mixed-effects model to assess internal consistency regarding the number of serves (ie, serving sizes of the food group consumed) reported. The intraclass correlation coefficient was 0.87, indicating good reliability. Cohen’s κ was calculated to evaluate agreement between test and retest scores. Agreement between the two tests was fair (κ=0.24, 95% CI=-0.15 to 0.63; P=0.239).

The IPAQ-C score was regarded as a categorical variable indicating exercise level based on the frequency and intensity of physical activity: (1) low (total activity <600 metabolic equivalent of task [MET]–minutes/week), (2) moderate (total activity ≥600 MET-minutes/week), or (3) high (total activity >3000 MET-minutes/week).

Finally, an unhealthy lifestyle score (0 to 4) was assigned to each participant based on a composite outcome involving low HKDS, physical inactivity, current smoking habit, and alcohol consumption; each unhealthy habit contributed 1 point to the score.

We used SPSS software (Windows version 26.0; IBM Corp, Armonk [NY], United States) for data analysis. Descriptive analyses were performed regarding the participants’ socio-demographic details, clinical information (eg, presence of chronic diseases), and the HKDS. The primary outcome variables included: (1) unhealthy dietary habits (low HKDS score); (2) suboptimal physical activity (low IPAQ-C score, indicating low exercise level); and (3) unhealthy lifestyle score (≥2). Univariable logistic regression was performed to examine associations between socio-demographic variables and each of the first two outcome variables. Multivariable logistic regression was modelled by controlling for covariates with P values <0.20 in univariable regression analysis, a cut-off level commonly used in public health research. For example, Torenfält and Dimberg13 utilised this approach when evaluating stroke and death in middle-aged Swedish men. The approach was also used in a French study14 concerning medical features of patients with COVID-19 and influenza. Additionally, linear regression analysis was conducted in the present study to examine associations between socio-demographic variables and the unhealthy lifestyle score. Time trends for various food intake, physical inactivity, current smoking, and alcohol consumption statuses were evaluated; the prevalences of these lifestyle habits were compared with population-wide figures from governmental reports over the past two decades using the Chi squared test for heterogeneity. P values <0.05 were considered statistically significant.

The Centre for Health Protection has been conducting health surveys periodically since 2003 to collect information about health and lifestyle-related behaviours, as well as practices related to the prevention of NCDs among residents aged ≥15 years.15 The resulting reports have presented key findings concerning physical activity, dietary habits, alcohol consumption, and smoking habits, as well as other self-care practices. We gathered relevant findings from 14 governmental reports covering the period from 2003-2004 to 2018-2019 (calendar years with the most updated figures)15 to perform trend analysis of fruit and vegetable consumption, physical activity level, smoking, and alcohol consumption status among Hong Kong residents. These results were compared with the findings of the present study; adjustments were solely performed for sex because the age distribution was limited in all but the most recent reports.

In total, 2551 individuals were contacted for a telephone interview and 1500 participated; the response rate was 58.8%. Most interviewed individuals were older adults (≥65 years, 66.7%), women (65.6%), and married (77.9%). Of the participants, 40.1% were engaged in professional and office work; only 16.3% had attained a tertiary education or higher. About 16% of participants reported a household income ≥HK$30 000, whereas 41.1% reported a household income <HK$10 000. Health status was predominately self-reported as average (42.5%) or above average (51.6%). More than half of the participants (54.5%) were undergoing long-term medical management or were taking medications for chronic diseases; the most common chronic conditions were diabetes mellitus (21.0%) and hypertension (42.7%) [Table 1].

Dietary habits, as measured by the HKDS (score range, 2-9), were classified as high scoring (5-9) or low scoring (0-4). Approximately 51% of participants had a low score (Table 1), suggesting minimal adherence to the traditional Chinese eating pattern; these participants were mainly younger individuals (aged ≤34 years, 66%) and men (58%) [Table 2]. Greater proportions of participants with lower income, current smokers, and current drinkers had low scores according to the HKDS (54%, 69%, and 67%, respectively) [Table 2]. Participants with chronic diseases had various HKDS results; <50% of patients with renal diseases and diabetes had a high score, and this result indicated that they had a poor dietary habits.

A greater risk of practising unhealthy dietary habits (low HKDS) was associated with male sex (adjusted odds ratio [aOR]=1.31, 95% CI=1.03-1.67), non-married status (ie, single/divorced/widowed) [aOR=1.56, 95% CI=1.20-2.03], a diagnosis of diabetes (aOR=1.53, 95% CI=1.15-2.03), and alcohol consumption (aOR=1.76, 95% CI=1.17-2.64) [Table 3].

Among all participants, 35.5%, 54.0% and 10.5% had low, moderate, and high levels of physical activity, respectively (Table 1). Participants who were non-married (aOR=1.66, 95% CI=1.23-2.22), undergoing long-term management of medical diseases (aOR=1.65, 95% CI=1.08-2.54), had diabetes (aOR=1.39, 95% CI=1.02 to 1.89), and had renal diseases (aOR=9.32, 95% CI=2.06-42.25) exhibited greater likelihood of physical inactivity (Table 4).

Few participants were current daily smokers (2.8%) and alcohol drinkers (8.5%) [Table 1].

Factors associated with an unhealthy lifestyle score are presented in Table 5. Male sex (beta coefficient [β]=0.25, 95% CI=0.16-0.34), non-married status (β=0.19, 95% CI=0.08-0.29), manual work (β=0.17, 95% CI=0.07-0.27), self-reported poor or very poor health status (β=0.26, 95% CI=0.07-0.45), a diagnosis of diabetes (β=0.31, 95% CI=0.20-0.41), and a diagnosis of renal disease (β=0.92, 95% CI=0.53-1.31) increased the likelihood of poor lifestyle habits. Housewife or retired status (β=-0.15, 95% CI=-0.25 to -0.04) and a higher household income (≥HK$30 000; β=-0.20, 95% CI=-0.33 to -0.08) decreased the likelihood of poor lifestyle habits.

The Health Behaviour Survey, with a response rate of 70.8% in the 2018/2019 report, is a population-based fieldwork study conducted by the Centre for Health Protection of the Department of Health.16 In that survey, female participants comprised 52.7% of the sample, compared with 65.6% in the present telephone survey. The age-group with the largest proportion of participants in the Survey was 65 to 74 years (36.7% vs 11.0%), which might have influenced the sex ratio.

Time trend analysis showed that the proportion of surveyed Hong Kong residents eating five daily servings of fruits and vegetables declined for both sexes in general (Fig a). Similarly, a significantly smaller proportion of participants reported walking >10 minutes for ≥5 days per week, and this proportion has continued to decline since 2016 (Fig b). There was a gradual decrease in the number of participants with a moderate or high level of physical activity. Despite a notable peak in 2019, there was a decline in 2020 with <60% of participants reportedly engaging in these physical activities. Finally, significantly smaller proportion of the study participants reported not currently smoking or consuming alcohol, compared with the proportions in previous population-based surveys (2010-2019) [Fig c and d].

In this population-based study of 1500 Hong Kong residents during the third wave of the COVID-19 pandemic, we found that the proportion of people with healthy food intake (ie, daily consumption of five servings of fruits and vegetables) has decreased since 2003; although a slight increased was observed in 2020, it was still below the overall average. Additionally, we found that the prevalence of low physical activity has gradually increased. In contrast, the rates of smoking and alcohol consumption were below the rates observed in pre-pandemic population-based surveys. Men and women in various age-groups had dietary habits less adherent to the traditional Chinese eating pattern, as measured by the HKDS, than 20 years prior. Adherence to the traditional eating pattern was significantly lower among male participants, single individuals, low-income participants, alcohol drinkers, individuals with low physical activity, and patients with diabetes mellitus or renal disease. We also found that men, individuals with non-married status, manual workers, individuals with self-perceived poor or very poor health status, and patients with diabetes and renal disease had a greater likelihood of poor lifestyle habits.

Worldwide, insufficient intake of fruits and vegetables and inadequate physical activity have been attributed to 34% and ≥20% of NCDs, respectively.17 18 The incidence of chronic diseases (eg, cancers, diabetes, and cardiovascular disease) in Hong Kong has increased by 60% in the past two decades.19 Although processed food intake was not assessed in this study, the HKDS results indicated very low intakes of legumes and red/orange vegetables among Hong Kong residents; a previous study showed that fruit intake in Hong Kong is among the lowest levels worldwide.20 The report of Health Behaviour Survey 2018/2019 revealed that 95.6% of surveyed participants had inadequate daily fruit and vegetable intake, based on World Health Organization recommendations.16 The present study showed further reduction, such that the proportion of Hong Kong residents consuming the recommended (≥5) daily servings of fruits and vegetables declined by 3.9% in the past two decades. Additionally, 10% of the population reported consuming more than one daily serving of processed meat, and the frequency of processed meat consumption increased by 3.1%.16 Another known risk factor for NCDs, overweight or obesity, affected approximately half of the Hong Kong population aged 15 to 84 years in 2015; this value was slightly (2%) below the global average.21

Disruptions to usual routines can lead to new dietary behaviours. For example, dine-in restrictions at restaurants and bars during the pandemic led to greater use of food delivery services, and take-away food is often less healthy.22 However, we observed increased fruit and vegetable consumption after the third wave of the COVID-19 outbreak. This finding is supported by the work of Wang et al,23 which showed that fruit and vegetable consumption increased. Collectively, the disease prevention and control policies that prohibited group gatherings and shortened the operating hours of bars and clubs (if such facilities were not entirely closed) could have significantly reduced smoking and alcohol consumption in Hong Kong during the pandemic. Additionally, the lower level of physical activity could be related to the closure of public sports avenues and restricted access to sports facilities.

The dietary habits of Hong Kong residents have changed from the traditional Chinese diet to a fast-paced dining experience involving convenient, processed foods with limited diversity. Hong Kong has a population of >7.5 million people, and 90% of its food supply is imported from other countries.24 Woo et al11 concluded that, despite geographical and cultural differences, traditional Chinese dietary habits were conceptually similar to the health-promoting Mediterranean diet. In 2001, high overall adherence to the Mediterranean diet was observed across all age-groups in Hong Kong, except for younger populations and men.11 In contrast, the present study showed that a smaller proportion of participants in all age and sex groups had high overall adherence to the Mediterranean diet. In this study, intriguingly, women consistently exhibited higher Mediterranean diet score (71%) and HKDS (53%; P<0.005) results, compared with their male counterparts. Moreover, the present study revealed lower prevalences of some chronic diseases among women: diabetes, cardiovascular disease, liver disease, and renal disease. The contemporary diets in modern Hong Kong and many developed regions have low fibre content and high processed food content; they also include food additives, refined sugar, and hydrogenated fats.25 The subtle but consistent westernisation of dietary habits appears to be detrimental for residents who consistently consume these food items.

A lack of colourful vegetables and fruits may reduce the diversity of beneficial gut microbiota,26 although leafy greens such as pak choy (Chinese cabbage), choy sum (Chinese flowering cabbage), and Chinese kale are staple foods in the Hong Kong diet throughout the year. Moreover, in a review of literature concerning exercise and gut microbial composition, Mitchell et al27 found that exercise alters gut microbiota; however, the direction of apparent change has varied among studies. Increases in butyrate-producing bacteria and faecal butyrate concentrations, with protective anti-inflammatory effects and the potential to enhance anti-infection immunity, have been observed among physically active adults.28 The mechanisms are unclear but the benefits of adopting a lifestyle with a diverse diet and physical activity consistently create an optimal environment for gut microbiota.

The large sample size and random sampling design of this territory-wide survey were strengths that enhanced the validity of the findings. However, this study had several limitations that should be addressed. First, cause-and-effect relationships between the COVID-19 pandemic and changes in lifestyle habits could not be established because of the cross-sectional approach. Individuals in home quarantine during the third wave of the pandemic might have experienced temporary changes in lifestyle habits; a prospective observational study and longer trend analysis are needed to facilitate long-term comparisons. Factors other than the pandemic (eg, mental wellness) could also have affected lifestyles among the study population. Although the present study utilised a random sampling strategy, non-response and selection biases were possible because younger segments of the Hong Kong population did not use landlines during the study period. Furthermore, response and social desirability biases may have been present in this telephone survey. Nevertheless, the high response rate and anonymous nature of this survey may have minimised these potential biases. Additionally, generalisation of the study findings should be performed with caution because the survey only included a single Chinese population. Considering that the participants’ characteristics differed from the general population, the findings might not be directly applicable to the general public. Moreover, the survey was conducted in 2020, and lifestyle habits in the general population might have changed throughout the pandemic. Finally, the low case numbers of some self-reported diseases, such as renal disease and cancer, may have resulted in type II error.

This representative population-based survey revealed that larger proportions of the general population had unhealthy lifestyles, including dietary habits and physical inactivity, during the COVID-19 pandemic than during pre-pandemic years. There is an urgent need to formulate and implement effective public health strategies at both individual and organisational levels. The encouragement of healthy lifestyles through evidence-based health promotion programmes is essential. This encouragement could be conveyed to communities through organised and concerted efforts by the government and relevant stakeholders. Future studies should evaluate the effectiveness of various interventions and approaches to achieve these important goals.

Concept or design: WYY Lin, MCS Wong.
Acquisition of data: WYY Lin, J Huang, Y Bai.
Analysis or interpretation of data: WYY Lin, J Huang, Y Bai.
Drafting of the manuscript: WYY Lin, MCS Wong.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

As editors of the journal, MCS Wong and J Huang were not involved in the peer review process. Other authors have disclosed no conflicts of interest.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This research was approved by the Survey and Behavioural Research Ethics Committee of The Chinese University of Hong Kong (Ref No.: SBRE-20-099). The interviewees provided informed consent after they were briefed on the study purpose and being assured of the confidentiality measures in place.

1. World Health Organization. WHO Coronavirus Disease (COVID-19) Dashboard. Available from: https://covid19.who.int/. Accessed 19 Jan 2024.

2. Centre for Health Protection, Department of Health, Hong Kong SAR Government. Situation of COVID-19 (23 January 2020 to 29 January 2023). 2023 July 27. Available from: https://www.chp.gov.hk/files/pdf/local_situation_covid19_en.pdf. Accessed 2 Feb 2024.

3. Chief Executive’s Office, Hong Kong SAR Government. Fighting the virus for six months battling another wave of the epidemic together. 2020 Jul 25. Available from: https://www.ceo.gov.hk/archive/5-term/eng/pdf/article20200725.pdf. Accessed 6 Apr 2022.

4. Mattioli AV, Sciomer S, Cocchi C, Maffei S, Gallina S. Quarantine during COVID-19 outbreak: changes in diet and physical activity increase the risk of cardiovascular disease. Nutr Metab Cardiovasc Dis 2020;30:1409-17. Crossref

5. Bin Mahfoz TM, Shaik RA, Ahmad MS, Alzerwi NA, Almutairi R. A nationwide survey to assess COVID-19’s impact on health and lifestyle in Saudi Arabia. Eur Rev Med Pharmacol Sci 2022;26:4092-101. Crossref

6. Mitchell JJ, Bu F, Fancourt D, Steptoe A, Bone JK. Longitudinal associations between physical activity and other health behaviours during the COVID-19 pandemic: a fixed effects analysis. Sci Rep 2022;12:15956. Crossref

7. Wang JX, Huang J, Cheung CS, Wong WN, Cheung NT, Wong MC. Adoption of an electronic patient record sharing pilot project: cross-sectional survey. J Med Internet Res 2020;22:e13761. Crossref

8. de Groot LC, van Staveren WA, Burema J. Survival beyond age 70 in relation to diet. Nutr Rev 1996;54:211-2. Crossref

9. Fung TT, Rexrode KM, Mantzoros CS, Manson JE, Willett WC, Hu FB. Mediterranean diet and incidence of and mortality from coronary heart disease and stroke in women. Circulation 2009;119:1093-100. Crossref

10. Bamia C, Lagiou P, Buckland G, et al. Mediterranean diet and colorectal cancer risk: results from a European cohort. Eur J Epidemiol 2013;28:317-28. Crossref

11. Woo J, Woo KS, Leung SS, et al. The Mediterranean score of dietary habits in Chinese populations in four different geographical areas. Eur J Clin Nutr 2001;55:215-20. Crossref

12. Centre for Food Safety, Hong Kong SAR Government. Nutrient and health—maintain optimal nutrient intake. 2008. Available from: https://www.cfs.gov.hk/english/multimedia/multimedia_pub/multimedia_pub_fsf_28_02.html. Accessed 30 Sep 2020. Crossref

13. Torenfält I, Dimberg L. Stroke and death — findings from a 25-year follow-up of a cohort of employed Swedish middle-aged men of the Coeur study. J Public Health 2022;30:1713-24. Crossref

14. Faury H, Courboulès C, Payen M, et al. Medical features of COVID-19 and influenza infection: a comparative study in Paris, France. J Infect 2021;82:e36-9. Crossref

15. Center for Health Protection, Department of Health, Hong Kong SAR Government. Non-communicable Disease Branch. Publications. Available from: https://www.chp.gov.hk/en/resources/29/100057.html. Accessed 31 Jan 2024.

16. Centre for Health Protection, Department of Health, Hong Kong SAR Government. Report of Health Behaviour Survey 2018/19. Available from: https://www.chp.gov.hk/files/pdf/report_of_health_behaviour_survey_2018_en.pdf. Accessed 29 Jan 2021.

17. World Health Organization. Global health risks: mortality and burden of disease attributable to selected major risks. 2009. Available from: https://apps.who.int/iris/handle/10665/44203. Accessed 29 Jan 2021.

18. World Health Organization. Physical activity. 2022. Available from: https://www.who.int/news-room/fact-sheets/detail/physical-activity. Accessed 22 Jan 2024.

19. Hospital Authority. Overview of Hong Kong Cancer Statistics of 2018. 2020. Available from: https://www3.ha.org.hk/cancereg/pdf/overview/Overview%20of%20HK%20Cancer%20Stat%202018.pdf. Accessed 30 Sep 2020.

20. Micha R, Khatibzadeh S, Shi P, et al. Global, regional and national consumption of major food groups in 1990 and 2010: a systematic analysis including 266 country-specific nutrition surveys worldwide. BMJ Open 2015;5:e008705. Crossref

21. World Health Organization. Obesity and overweight. 2021. Available from: https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight. Accessed 23 Jun 2021.

22. Nguyen TH, Vu DC. Food delivery service during social distancing: proactively preventing or potentially spreading coronavirus disease-2019? Disaster Med Public Health Prep 2020;14:e9-10. Crossref

23. Wang J, Yeoh EK, Yung TK, et al. Change in eating habits and physical activities before and during the COVID-19 pandemic in Hong Kong: a cross-sectional study via random telephone survey. J Int Soc Sports Nutr 2021;18:33. Crossref

24. Research Office, Legislative Council Secretariat. Fact sheet: regulation of imported food in Hong Kong. 2016. Available from: https://www.legco.gov.hk/research-publications/english/1516fsc14-regulation-of-imported-food-in-hong-kong-20160226-e.pdf. Accessed 22 Jan 2024.

25. Jew S, AbuMweis SS, Jones PJ. Evolution of the human diet: linking our ancestral diet to modern functional foods as a means of chronic disease prevention. J Med Food 2009;12:925-34. Crossref

26. Garcia-Mantrana I, Selma-Royo M, Alcantara C, Collado MC. Shifts on gut microbiota associated to Mediterranean diet adherence and specific dietary intakes on general adult population. Front Microbiol 2018;9:890. Crossref

27. Mitchell CM, Davy BM, Hulver MW, Neilson AP, Bennett BJ, Davy KP. Does exercise alter gut microbial composition? a systematic review. Med Sci Sports Exerc 2019;51:160-7. Crossref

28. Allen JM, Mailing LJ, Niemiro GM, et al. Exercise alters gut microbiota composition and function in lean and obese humans. Med Sci Sports Exerc 2018;50:747-57. Crossref

Vaccination is an effective tool to combat the coronavirus disease 2019 (COVID-19) pandemic. Two types of COVID-19 vaccines are used in Hong Kong: the Sinovac-CoronaVac inactivated severe acute respiratory syndrome coronavirus 2 vaccine (Sinovac Biotech Ltd, Beijing, China) and Pfizer BioNTech BNT162b2 (Pfizer Inc, Philadelphia [PA], United States) messenger RNA vaccine began distribution on 26 February 2021 and 10 March 2021, respectively.

According to the World Health Organization, vaccine hesitancy is defined as delaying or refusing vaccination despite the availability of vaccination services.1 In a study conducted during the third wave of COVID-19 in Hong Kong, the overall vaccine acceptance rate was approximately 37%.2 Although the subsequent acceptance rate has varied with pandemic progression, confidence in COVID-19 vaccines has remained a key factor in reducing vaccine hesitancy.3

Pregnant women were generally excluded from clinical trials focusing on the development, safety, and efficacy of COVID-19 vaccines.4 When COVID-19 vaccines were introduced in Hong Kong, routine vaccination was not recommended for women who were pregnant or breastfeeding, except when there was a high risk of exposure or complications.5 The relative lack of data may have contributed to vaccine hesitancy among pregnant women.6 7 8 Based on data concerning the efficacy and safety of COVID-19 vaccination in preventing serious illness,9 COVID-19 vaccination is recommended for people who are pregnant, breastfeeding, planning to become pregnant, or may become pregnant in the future.10 11

On 23 April 2021, the Hong Kong College of Obstetricians and Gynaecologists (HKCOG) issued an interim recommendation that pregnant women receive the BioNTech COVID-19 vaccine at the same time as the general population.12 On 18 February 2022, the Sinovac vaccine was also recommended for use in pregnant women.12 Furthermore, the recommended interval between the second and third doses of COVID-19 vaccine was shortened from 180 days to 90 days, beginning on 4 March 2022. The vaccine pass policy for entry to specific premises was tightened on 31 May 2022.13 For persons who were over 18 years old and had no history of infection, a minimal of two doses of vaccination was required. A third dose was required if the second dose was taken over 6 months ago. Starting from 13 June 2022, women attending obstetric clinics were required to provide a negative result proof of a polymerase chain reaction–based nucleic acid test conducted with specimen collected within 48 hours before the visit if they did not fulfil the vaccine pass requirement.14

The fifth wave of COVID-19 in Hong Kong has resulted in an overwhelming number of COVID-19 cases. Pregnant women are not automatically protected from COVID-19; indeed, their vaccine hesitancy and low vaccination rate may lead to greater susceptibility. Information concerning the patterns of COVID-19 vaccination and disease transmission among pregnant women in Hong Kong is unavailable. This study examined patterns of vaccination and transmission among pregnant women who were booked for delivery at a tertiary hospital in Hong Kong, with the goal of providing insights into maternal disease characteristics.

This retrospective review included women who were booked for delivery at Queen Mary Hospital in Hong Kong and had attended the booking antenatal visit from 1 July 2021 to 31 March 2022. Information concerning COVID-19 vaccination history was retrieved from the Clinical Management System of the Hospital Authority, which captured COVID-19 vaccination data from the Department of Health.

Pregnant women were diagnosed with COVID-19 because of symptoms or (in the absence of symptoms) during admission screening. Women diagnosed with COVID-19 through other channels were able to reschedule their appointments. Phone consultations were provided by the obstetric team at Queen Mary Hospital. The clinical details of COVID-19 cases were documented in the computerised Antenatal Record System. Additionally, antenatal progress notes were updated if a pregnant woman reported a history of COVID-19 during a follow-up visit. Data regarding demographic characteristics, COVID-19 status, and obstetric outcomes were retrieved from the Antenatal Record System.

The vaccinated group comprised women who received at least one dose of any type of COVID-19 vaccine. Vaccination periods were classified as pre-pregnancy, antenatal, and postnatal for women with a known date of delivery, miscarriage, or termination of pregnancy as of 30 June 2022. For women with ongoing pregnancies and unknown obstetric outcomes, the vaccination period was estimated according to the expected date of delivery. Antenatal status was regarded as known ongoing pregnancy before 42 weeks of gestation. A vaccination episode was defined as any episode of COVID-19 vaccination including the first, second, and third doses. The number of days elapsed since vaccination was defined as the interval between the last dose of COVID-19 vaccine and 30 June 2022 for women who had received one or two doses of vaccine. Complete vaccination was regarded as the period between 15 and 90 days after the second dose of COVID-19 vaccine, or 14 days after the third dose of COVID-19 vaccine for women who had never been diagnosed with COVID-19. For women with COVID-19, the date of diagnosis was regarded as the reference point when determining vaccination status.

Descriptive statistics were reported. Vaccination rates were calculated according to age-group. Background demographic characteristics were compared between vaccinated and unvaccinated groups. Student’s t test, analysis of variance, and the Chi squared test were used as appropriate. Regression analyses were conducted to identify factors affecting vaccine acceptance. P values <0.05 were considered statistically significant. Statistical analyses were performed using SPSS software (Windows version 26; IBM Corp, Armonk [NY], United States).

Table 1 shows the demographic characteristics of 2396 pregnant women who had attended the booking antenatal visit between 1 July 2021 and 31 March 2022. As of 30 June 2022, 2006 (83.7%), 1843 (76.9%), and 831 (34.7%) women had received the first, second, and third doses of COVID-19 vaccine, respectively. Among the 1843 women who had received two doses of vaccine, 1056 (57.3%) underwent vaccination before pregnancy (Fig 1). Of these 1843 women, 831 received a third dose; the median interval between the second and third doses was 280 days (interquartile range, 239-308). Of the remaining 1012 women who had received only two doses of vaccine, 684 (67.6%) and 504 (49.8%) had already passed the 90-day and 180-day intervals, respectively. Their median number of days elapsed since the last vaccine was 315 (interquartile range, 145-368), which considerably exceeded the recommended 90-day interval.

Only 26.6% (1243/4680) of vaccination episodes occurred during pregnancy. Among women who underwent antenatal vaccination, 65.7% (817/1243) had it during the fifth wave of COVID-19 between January 2022 and June 2022; 65.7% (537/817) of these women received the third dose. The two peaks of vaccination for third dose were observed in early March 2022 and late May 2022 (Fig 2).

The vaccination rate was the lowest among Chinese women (81.4%), but the highest among Caucasian women (96.4%) [Fig 3]. Multivariate analysis showed that active working status (odds ratio [OR]=1.94; 95% confidence interval [CI]=1.47-2.56) was significantly associated with a higher COVID-19 vaccination rate, whereas Chinese ethnicity (OR=0.21; 95% CI=0.13-0.33) and women with obstetric complications (OR=0.72; 95% CI=0.55-0.94) were significantly associated with a lower COVID-19 vaccination rate.

In total, there were 265 (11.1%) COVID-19 cases in this cohort; the earliest diagnosis was made on 1 January 2022 during the fifth wave of COVID-19 in Hong Kong (Table 2). The disease rate was more than tenfold higher in women who had no (20.3%) or incomplete (18.8%) vaccination, compared with women who had complete vaccination (2.1%; P<0.001). After exclusion of pregnancies among women aged ≤19 years, there was an insignificant trend of lower vaccination among young women (ie, aged 20-29 years), but their disease rate was the highest (19.1%) [Fig 4].

Among 237 (89.4%) women who had an antenatal diagnosis of COVID-19, 42 (17.7%) required admission for monitoring and 26 (11.0%) delivered in an isolation facility. No women required intensive care or oxygen support. There were no adverse maternal outcomes or cases of vertical transmission.

To our knowledge, this is the first report of the low COVID-19 vaccination rate (83.7%) among pregnant women in Hong Kong. This rate is considerably lower than the single-dose rate among the general public (92.7%) that was reported on the government’s vaccination dashboard on the final day of the study period (ie, 30 June 2022).15 Furthermore, the disease rate was more than threefold higher in women who had no or incomplete vaccination, compared with women who had complete vaccination.

Pregnant women are considered a vulnerable group. The substantial increase in the disease rate, combined with a lower vaccination rate, among women aged 20 to 29 years is particularly concerning. A case of COVID-19 during pregnancy can lead to adverse obstetric outcomes, including increased risks of preterm delivery, growth restriction, and stillbirth.16 In addition to the effectiveness of vaccination in terms of reducing severe complications, the transplacental transfer of immunoglobulins after maternal vaccination might provide infants with protection against COVID-19.10 16 17 18

Vaccine hesitancy is a potential public health problem.19 The five psychological antecedents of vaccination are confidence, complacency, constraints, calculation, and collective responsibility.20 21 Acceptance of COVID-19 vaccination has varied among countries, with the highest rates reported in India, the Philippines, and Latin America.22 In the present study, Chinese women had a lower vaccination rate, compared with their non-Chinese counterparts.

The COVID-19 pandemic has contributed to a decrease in vaccine hesitancy.23 In May 2021, a systemic review showed that an estimated 47% of pregnant women worldwide intended to undergo COVID-19 vaccination.24

Communities generally become more complacent when the number of disease cases is low, suggesting that the perceived risk of disease transmission is minimal. This phenomenon was evident in Hong Kong across several waves of COVID-19 transmission.15 The vaccination rate increased during the fifth wave of COVID-19 when there was an exponential surge in the number of disease cases. A similar pattern was observed in the present study, such that two-thirds of the antenatal vaccination episodes occurred during the fifth wave of COVID-19 in Hong Kong.

Despite this relative surge in vaccination, around two-thirds of women eligible for the third dose did not receive it during pregnancy. This delay poses a major threat because the protective effect of the previous two doses may have dissipated. Importantly, although the rate of vaccination was higher in the antenatal group than the postnatal group, most women in the antenatal group had undergone vaccination before pregnancy. Thus, they may have a higher risk of serious adverse effects from COVID-19 if they become ill in the peripartum period. After exclusion of the small number of pregnant women aged ≤19 years, vaccination rates were consistently low among pregnant women of all ages. Pregnant women may have a higher risk of disease transmission in future waves of COVID-19; at the end of the present study, only one-third of pregnant women in this cohort had received three doses of vaccine.

Lack of confidence has been identified as the main factor consistently associated with lower COVID-19 vaccine hesitancy.3 The implementation of the stringent vaccine pass policy had driven another peak of vaccination in late May 2022. However, nearly half of the women who had received two doses of vaccine were indeed overdue for the third dose, and hence did not fulfil the vaccine pass for entrance to specific premises or could not attend obstetric clinics without undergoing nucleic acid testing. This group of vaccine had received COVID-19 vaccine before but had it mostly before pregnancy. Pregnancy could represent the key hindrance for their vaccination in the antenatal period. Government policy might not be adequate enough for promoting vaccination in pregnant women. Concern about possible harmful side-effects was the top reason for reluctance; confidence in COVID-19 vaccine safety and efficacy was the main predictor of acceptance, particularly in the pregnant population.8 22 In a Japanese cohort, concerns about potential effects on the fetus and breastfeeding were the main reasons for low COVID-19 vaccination acceptance.25 These findings highlight the need to distribute correct information and provide sufficient education to address concerns among women of reproductive age. In particular, antenatal women should receive additional information concerning vaccine safety during pregnancy.

A study in Hong Kong showed that recommendations from the government constituted the strongest factor driving COVID-19 vaccine acceptance.2 Education about the safety and benefit of vaccination is also important.7 Webinars would be useful in efforts to educate the general public. Within the hospital setting, vaccination teams comprising obstetricians and midwives could allay concerns and dispel myths about vaccination among working staff at all levels; this approach could provide useful information for pregnant and breastfeeding women. There is also a need to combat physician hesitancy in recommending COVID-19 vaccination for pregnant women.26 To address this need, the HKCOG revised its recommendations on 3 March 2022 to indicate that women who are planning to become pregnant, are pregnant, or are breastfeeding should undergo COVID-19 vaccination along with the general population.12 A corresponding educational video to promote COVID-19 vaccination was made available on 5 March 2022.12

Possible interventions to promote vaccination in Hong Kong include the provision of vaccines at convenient venues and the involvement of healthcare professionals in information dissemination.27 During the fifth wave of COVID-19, pregnant women were proactively asked to consider COVID-19 vaccination when they attended obstetric clinics. Leaflets were distributed with information about the HKCOG recommendations, as well as community vaccination sites. The establishment of a pathway specifically for pregnant women, which reduced their waiting time in vaccine clinics, also helped to increase the vaccination rate. Moreover, vaccination was provided to women in the maternity wards of some hospitals and women attending antenatal clinics in Maternal and Child Health Centres. All of these measures helped reduce vaccine hesitancy in pregnant women.28

In Hong Kong, a pertussis vaccination programme for pregnant women was launched on 2 July 2020.29 The programme was incorporated into antenatal care, such that all pregnant women received counselling concerning the rationale for vaccination; the vaccine was administered during obstetric follow-up. To facilitate vaccine availability in our hospital, all women were asked to indicate their preference concerning pertussis vaccination during the first antenatal visit. It may be useful to incorporate COVID-19 vaccination into the maternal immunisation programme.

This study had some notable strengths. The results of the present large cohort study provide clinicians and policymakers with key insights concerning the COVID-19 vaccination rate among pregnant women in Hong Kong. The study period was designed to include both antenatal and postnatal periods for a better understanding of vaccination behaviour among women in each period. A real-time collection method was adopted to capture COVID-19 vaccination records from the Clinical Management System, thereby ensuring data reliability.

Nevertheless, this study had some limitations. A small number of women who underwent COVID-19 vaccination outside Hong Kong were not automatically identified in the system; however, they were included in the cohort if their vaccination history had been documented in antenatal records. Because rapid antigen self-tests were acceptable for diagnosis in Hong Kong beginning on 26 February 2022, the number of recorded COVID-19 cases in our cohort might have been lower than the actual number of cases if patients did not report positive COVID-19 test results to our department. Finally, this study did not explore whether the generally more cautious approach of pregnant women, in terms of avoiding all types of diseases, might contribute to a lower disease rate compared with the general public.

The rate of COVID-19 vaccination was low among pregnant and postnatal women in Hong Kong in early 2022. Pregnant women had a high risk of disease transmission because many of them had not received the third dose of COVID-19 vaccine. Urgent measures are needed to promote vaccination among pregnant women before future waves of COVID-19. In particular, women should receive information concerning vaccine safety to avoid unnecessary delays related to pregnancy.

Concept or design: PW Hui, DTY Chan.
Acquisition of data: All authors.
Analysis or interpretation of data: PW Hui.
Drafting of the manuscript: PW Hui, LM Yeung, JKY Ko, THT Lai, MTY Seto.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

The authors thank Dr TW Chay, Dr YF Chiu, Ms WK Choi, Dr TY Hui, Dr KH Lam, Dr KS Lee, Dr YH Luk, Dr SK Ma, Dr HS Wang, Dr CCY Wong, Dr CL Wong, Dr LC Wong, and Dr CWK Yan of Department of Obstetrics and Gynaecology, Queen Mary Hospital for their contributions to research data retrieval and entry.

The abstract has been presented online in the Royal College of Obstetricians and Gynaecologists World Congress 2023 (14 June 2013, London, United Kingdom).

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Ethics approval was obtained from the Institutional Review Board of The University of Hong Kong/Hospital Authority Hong Kong West Cluster (Ref No.: UW 22-205). Informed patient consent has been waived by the Board due to the retrospective nature of the research.

1. MacDonald NE; SAGE Working Group on Vaccine Hesitancy. Vaccine hesitancy: definition, scope and determinants. Vaccine 2015;33:4161-4. Crossref

2. Wong MC, Wong EL, Huang J, et al. Acceptance of the COVID-19 vaccine based on the health belief model: a population-based survey in Hong Kong. Vaccine 2021;39:1148-56. Crossref

3. Xiao J, Cheung JK, Wu P, Ni MY, Cowling BJ, Liao Q. Temporal changes in factors associated with COVID-19 vaccine hesitancy and uptake among adults in Hong Kong: serial cross-sectional surveys. Lancet Reg Health West Pac 2022;23:100441. Crossref

4. Beigi RH, Krubiner C, Jamieson DJ, et al. The need for inclusion of pregnant women in COVID-19 vaccine trials. Vaccine 2021;39:868-70. Crossref

5. Scientific Committee on Emerging and Zoonotic Disease and Scientific Committee on Vaccine Preventable Diseases, Centre for Health Protection, Hong Kong SAR Government. Consensus interim recommendations on the use of COVID-19 vaccines in Hong Kong (as of Jan 7, 2021). Available from: https://www.chp.gov.hk/files/pdf/consensus_interim_recommendations_on_the_use_of_covid19_vaccines_inhk.pdf. Accessed 30 Mar 2022.

6. Tao L, Wang R, Han N, et al. Acceptance of a COVID-19 vaccine and associated factors among pregnant women in China: a multi-center cross-sectional study based on health belief model. Hum Vaccin Immunother 2021;17:2378-88. Crossref

7. Sznajder KK, Kjerulff KH, Wang M, Hwang W, Ramirez SI, Gandhi CK. COVID-19 vaccine acceptance and associated factors among pregnant women in Pennsylvania 2020. Prev Med Rep 2022;26:101713. Crossref

8. Sutton D, D’Alton M, Zhang Y, et al. COVID-19 vaccine acceptance among pregnant, breastfeeding, and nonpregnant reproductive-aged women. Am J Obstet Gynecol MFM 2021;3:100403. Crossref

9. Shimabukuro TT, Kim SY, Myers TR, et al. Preliminary findings of mRNA COVID-19 vaccine safety in pregnant persons. N Engl J Med 2021;384:2273-82. Crossref

10. Halasa NB, Olson SM, Staat MA, et al. Effectiveness of maternal vaccination with mRNA COVID-19 vaccine during pregnancy against COVID-19–associated hospitalization in infants aged <6 months—17 states, July 2021–January 2022. MMWR Morb Mortal Wkly Rep 2022;71:264-70. Crossref

11. Girardi G, Bremer AA. Scientific evidence supporting coronavirus disease 2019 (COVID-19) vaccine efficacy and safety in people planning to conceive or who are pregnant or lactating. Obstet Gynecol 2022;139:3-8. Crossref

12. The Hong Kong College of Obstetricians and Gynaecologists. The Hong Kong College of Obstetricians and Gynaecologists advice on COVID-19 vaccination in pregnant and lactating women (interim; updated on 3rd March 2022). 2022. Available from: http://www.hkcog.org.hk/hkcog/Upload/EditorImage/20220304/20220304134806_9035.pdf. Accessed 30 Mar 2022.

13. Hong Kong SAR Government. FEHD reminds catering premises operators and customers to observe requirements on third stage of Vaccine Pass. 2022. Available from: https://www.info.gov.hk/gia/general/202205/30/P2022053000726.htm?fontSize=1. Accessed 30 May 2022.

14. Hong Kong SAR Government. Government to implement Vaccine Pass arrangement in designated healthcare premises. 2022. Available from: https://www.info.gov.hk/gia/general/202205/21/P2022052100421.htm. Accessed 21 May 2022.

15. Together, We Fight the Virus. The Government of the Hong Kong Special Administrative Region. Available from: https://www.coronavirus.gov.hk/eng/index.html. Accessed 30 Jun 2022.

16. Allotey J, Stallings E, Bonet M, et al. Clinical manifestations, risk factors, and maternal and perinatal outcomes of coronavirus disease 2019 in pregnancy: living systematic review and meta-analysis. BMJ 2020;370:m3320. Crossref

17. Male V. SARS-CoV-2 infection and COVID-19 vaccination in pregnancy. Nat Rev Immunol 2022;22:277-82. Crossref

18. Nir O, Schwartz A, Toussia-Cohen S, et al. Maternal-neonatal transfer of SARS-CoV-2 immunoglobulin G antibodies among parturient women treated with BNT162b2 messenger RNA vaccine during pregnancy. Am J Obstet Gynecol MFM 2022;4:100492. Crossref

19. Shook LL, Kishkovich TP, Edlow AG. Countering COVID-19 vaccine hesitancy in pregnancy: the “4 Cs”. Am J Perinatol 2022;39:1048-54. Crossref

20. Betsch C, Schmid P, Heinemeier D, Korn L, Holtmann C, Böhm R. Beyond confidence: development of a measure assessing the 5C psychological antecedents of vaccination. PLoS One 2018;13:e0208601. Crossref

21. Chau CY. COVID-19 vaccination hesitancy and challenges to mass vaccination. Hong Kong Med J 2021;27:377-9. Crossref

22. Skjefte M, Ngirbabul M, Akeju O, et al. COVID-19 vaccine acceptance among pregnant women and mothers of young children: results of a survey in 16 countries. Eur J Epidemiol 2021;36:197-211. Crossref

23. Gencer H, Özkan S, Vardar O, Serçekuş P. The effects of the COVID-19 pandemic on vaccine decisions in pregnant women. Women Birth 2022;35:317-23. Crossref

24. Shamshirsaz AA, Hessami K, Morain S, et al. Intention to receive COVID-19 vaccine during pregnancy: a systematic review and meta-analysis. Am J Perinatol 2022;39:492-500. Crossref

25. Hosokawa Y, Okawa S, Hori A, et al. The prevalence of COVID-19 vaccination and vaccine hesitancy in pregnant women: an internet-based cross-sectional study in Japan. J Epidemiol 2022;32:188-94. Crossref

26. Chervenak FA, McCullough LB, Grünebaum A. Reversing physician hesitancy to recommend COVID-19 vaccination for pregnant patients. Am J Obstet Gynecol 2022;226:805-12. Crossref

27. Wang K, Wong EL, Cheung AW, et al. Influence of vaccination characteristics on COVID-19 vaccine acceptance among working-age people in Hong Kong, China: a discrete choice experiment. Front Public Health 2021;9:793533. Crossref

28. Iacobucci G. COVID-19 and pregnancy: vaccine hesitancy and how to overcome it. BMJ 2021;375:n2862. Crossref

29. Hospital Authority. Public hospital pertussis vaccination programme for pregnant women. Jun 28, 2020. Available from: https://www.ha.org.hk/haho/ho/cc/pertussis_press_release_en.pdf. Accessed 29 Jun 2023.

Shoulder dystocia is an uncommon obstetric emergency with an incidence that reportedly ranging from 0.58% to 0.7%.1 2 3 It can result in severe perinatal morbidities, including brachial plexus palsies, clavicular fractures, humeral fractures, hypoxic-ischaemic encephalopathy, cerebral palsy, and even mortality soon after birth.1 2 3 4 5 6 Considering the unpredictable and complex nature of shoulder dystocia, many professional bodies have established systematic approaches with routine training simulations and algorithms to improve fetal outcomes in such cases.1 2 7 8 9 The most common approach is represented by the HELPERR mnemonic, which consists of a sequence of manoeuvres including the McRoberts’ manoeuvre, suprapubic pressure, rotational methods, posterior arm delivery, all-fours position, and clavicular fracture.7 Although the McRoberts’ manoeuvre and suprapubic pressure are often the preferred initial manoeuvres, their success rates (56.0%) are lower than that of the rotational methods (62.4%) and posterior arm delivery (86.1%).5 6 10 The Royal College of Obstetricians and Gynaecologists (RCOG) in 20121 and the American College of Obstetricians and Gynecologists in 20172 revised their guidelines to indicate that either posterior arm delivery or rotational methods can be used after an unsuccessful attempt of the McRoberts’ manoeuvre. Furthermore, a randomised controlled trial published in 2015 demonstrated that the induction of labour at 38 weeks in macrosomic pregnancies could reduce the risk of shoulder dystocia, compared with expectant management.11 The overall caesarean rate did not increase when using this approach. Accordingly, early induction of labour has become an option in cases of suspected fetal macrosomia.

Considering changes in the management of macrosomic pregnancies and shoulder dystocia in the past decade, this study compared the incidences of shoulder dystocia, the maternal and fetal characteristics in such cases, and their obstetric management methods and perinatal outcomes between the periods of 2000-2009 and 2010-2019.

This retrospective study was conducted in a university tertiary obstetric unit that provided obstetric services in the New Territories East Cluster of Hong Kong. All consecutive cases of shoulder dystocia reported from 2000 to 2019 inclusive were identified using the hospital’s electronic database.12 Shoulder dystocia was objectively defined as the requirement of an ancillary obstetric manoeuvre following failed delivery of the anterior shoulder after downward fetal neck traction or head-to-body delivery interval (HBDI) >1 minute, as described in previous reports.4 5 13 Multiple pregnancies, vaginal breech deliveries, and known stillbirths before labour were excluded. Our unit protocol for the management of shoulder dystocia followed the RCOG Green-top Guidelines for shoulder dystocia that was published in 2005 and updated in 2012.1 Beginning in 2002, hands-on training in shoulder dystocia relief was routinely conducted using the ALSO (Advanced Life Support in Obstetrics) program7; the PROMPT (Practical Obstetric Multi-Professional Training)8 and the SOPHIE (Safe Obstetric Practice for High risk and Emergency)9 training methods were added in 2011, after the publication of our articles regarding shoulder dystocia.4 5 The McRoberts’ manoeuvre with or without suprapubic pressure was usually the first manoeuvre performed, followed by a rotational manoeuvre or posterior arm delivery if the McRoberts’ manoeuvre was unsuccessful. A midwife was designated to document each event, including the personnel involved, usage and duration of manoeuvres, and delivery times of the fetal head and fetal body. Umbilical cord blood was collected immediately after delivery for blood gas analysis using a Bayer Rapidpoint 400 Blood Gas Analyzer (Bayer HealthCare, Seattle [WA], United States), as described in our previous reports.14 15 Delivery data, including birth weight and perinatal complications, were recorded immediately after delivery by the attending staff, then crosschecked by another staff member. Two to 3 months later, fetal outcomes were subjected to further review and confirmation during postnatal follow-up and monthly audit meetings. All midwives and obstetricians attended annual training sessions regarding the management of shoulder dystocia.

Identified cases and corresponding medical records were reviewed to collect maternal, fetal, and obstetric characteristics. Advanced maternal age was defined as ≥35 years at the estimated date of delivery. Short stature was defined as maternal height <150 cm. Maternal body weights at booking and delivery were recorded to calculate the body mass index (BMI) at each time point. Obesity was defined as BMI >30 kg/m², in accordance with World Health Organization guidelines.16 17 Parity and diabetes mellitus/gestational diabetes mellitus statuses were noted. Obstetric and neonatal characteristics were recorded, including delivery mode, fetal distress, birth weight, HBDI duration, and the use and sequence of manoeuvres. Macrosomia was defined as birth weight ≥4000 g. Neonatal outcomes were also recorded, including the Apgar scores at 1 minute and 5 minutes of life, cord arterial pH, and neonatal complications (eg, subgaleal haemorrhage, hypoxic-ischaemic encephalopathy, brachial plexus injury, clavicular fracture, and humeral fracture).

The incidences of shoulder dystocia were calculated for three groups: all singleton live pregnancies (excluding multiple pregnancies and stillbirths), all singleton live cephalic-presenting pregnancies with spontaneous onset of labour (excluding cases of fetal malpresentation and elective caesarean deliveries), and all singleton live cephalic-presenting pregnancies with successful vaginal delivery (excluding emergency caesarean deliveries).12 The incidences of shoulder dystocia were also calculated for various birth weight ranges.

The incidences of shoulder dystocia, maternal and fetal characteristics in each case, and perinatal outcomes were compared between 2000-2009 and 2010-2019. Analyses were performed using SPSS (Windows version 26.0; IBM Corp, Armonk [NY], United States). The incidences of shoulder dystocia; maternal, obstetric, and fetal characteristics; and neonatal complications related to shoulder dystocia during 2000-2009 and 2010-2019 were analysed using the Chi squared test and Fisher’s exact test for categorical variables, t test for parametric continuous variables, and Mann-Whitney U test for non-parametric continuous variables. The threshold for statistical significance was set at P<0.05.

In total, this study included 242 cases of shoulder dystocia in the study unit. Table 1 shows the incidences of shoulder dystocia in each decade according to type of birth and range of birth weight. The overall incidence of shoulder dystocia among all singleton live pregnancies decreased from 0.23% (134/58 326) during 2000-2009 to 0.16% (108/65 683) during 2010-2019 (P=0.009). Similarly, the incidence of shoulder dystocia among singleton cephalic-presenting pregnancies with spontaneous onset of labour decreased from 0.25% to 0.19% (P=0.031), and the incidence of shoulder dystocia among singleton cephalic-presenting pregnancies with vaginal delivery decreased from 0.29% to 0.21% (P=0.017). The incidences of shoulder dystocia were generally similar across birth weight categories, but the incidence considerably decreased in the 4200-4399 g group.

Table 2 shows the birth weight distribution for all singleton cephalic live pregnancies. The mean birth weight decreased from 3180 ± 472 g during 2000-2009 to 3132 ± 463 g during 2010-2019 (P<0.001). The proportion of babies weighing ≥4000 g was 3.3% during 2000-2009, whereas it was 2.3% during 2010-2019.

Table 3 illustrates the maternal and obstetric characteristics in cases of shoulder dystocia during each decade. There were no statistically significant differences between the two decades in terms of advanced maternal age, maternal age, maternal height, maternal weight at booking and delivery, BMI at booking and delivery, obesity at delivery, or nulliparity. However, the proportion of shoulder dystocia cases involving maternal diabetes increased from 9.0% during 2000-2009 to 19.4% during 2010-2019 (P=0.018). Additionally, the proportion of shoulder dystocia cases involving instrumental delivery decreased from 65.7% to 47.2% (P=0.004), but there was no statistically significant difference in the proportion of deliveries involving fetal distress. Although there was no significant change in birth weight, the proportion of babies with macrosomia among shoulder dystocia cases tended to decrease over time (from 27.6% to 18.5%; P=0.097).

Table 4 shows the success rates of various manoeuvres in terms of alleviating shoulder dystocia. During 2010-2019, 79.6% of babies in cases of shoulder dystocia had HBDI ≤2 minutes; 14.8% and 5.6% of such babies had HBDI of 3-4 minutes and ≥5 minutes, respectively. These proportions were significantly better than the proportions during 2000-2009 (59.0%, 31.3%, and 9.7%, respectively; P=0.003). The success rate of the McRoberts’ manoeuvre in terms of alleviating shoulder dystocia increased from 31.3% to 47.2% (P=0.012) among all vaginal deliveries, which was partially attributed to the increased success rate among instrumental deliveries (from 20.5% to 39.2%; P=0.017). Although the rotational manoeuvre continued to be preferred over posterior arm extraction (77.6% vs. 22.4%) after failure of the McRoberts’ manoeuvre and suprapubic pressure, the success rate of posterior arm extraction increased from 52.9% in 2000-2009 to 92.3% in 2010-2019 (P=0.042). Table 5 shows the neonatal outcomes of shoulder dystocia. There were significant reductions in the rates of low Apgar scores: for an Apgar score <5 at 1 minute of life, the rate decreased from 13.4% to 5.6% (P=0.042); for an Apgar score <7 at 5 minutes of life, the rate decreased from 11.9% to 4.6% (P=0.045). There were no statistically significant changes in the rates of other neonatal complications.

This study revealed a significant reduction in the overall incidence of shoulder dystocia over the past two decades in a tertiary obstetric unit in Hong Kong. One possible reason is the increased use of caesarean delivery, in both elective and emergency settings, in cases of suspected macrosomia. This hypothesis is supported by the decrease in the incidences of shoulder dystocia among all births and among all pregnancies with onset of labour in the 4200-4399 g subgroup during 2010-2019 (Table 1). However, such decreases were not observed in other subgroups with babies weighing ≥4000 g (4000-4199 g, 4400-4599 g, and ≥4600 g) [Table 1]. These findings suggest that caesarean delivery in cases of suspected macrosomia is not a major factor contributing to shoulder dystocia prevention. Furthermore, the absence of any change in shoulder dystocia incidence among subgroups with babies weighing <4200 g was consistent with our departmental practice of using 4200 g, rather than 4000 g, as a threshold for offering caesarean delivery to non-diabetic women. The use of a lower fetal weight threshold (eg, 4000 g) in pregnant Chinese women could lead to an unnecessary increase in the rate of caesarean delivery.18 The practice of early induction of labour in cases of suspected macrosomia may be the main factor contributing to the decrease in shoulder dystocia incidence.11 Since 2011, women with a fetal abdominal circumference or estimated fetal weight above the 97th percentile (but <4200 g) have been counselled about the risk of difficult labour, along with the benefits and risks of inducing labour to prevent further macrosomia. These approaches are consistent with the decreases in overall mean birth weight and proportion of babies ≥4000 g from 2000-2009 to 2010-2019 (Table 2).

Additionally, our study revealed improvements in the acute management of shoulder dystocia. In particular, the overall success rate of the McRoberts’ manoeuvre in terms of alleviating shoulder dystocia improved from 31.3% during 2000-2009 to 47.2% during 2010-2019; these results were consistent with rates in published reports, which have considerably varied from 23% to 70%.5 6 19 20 21 Our improved success rate was mainly attributed to the increased success rate among instrumental deliveries (from 20.5% to 39.2%) [Table 4]. We previously speculated that instrumental delivery increased the risk of shoulder dystocia while reducing the likelihood of McRoberts’ manoeuvre success, presumably related to delayed descent of the shoulders during instrumental delivery.22 The SOPHIE training emphasises that proper performance of the McRoberts’ manoeuvre should result in cephaloid rotation of the mother’s pelvis, manifested by elevation of the mother’s buttocks from the bed. To achieve this goal, the best method for hyperflexion of the mother’s hips involves grasping the back of the mother’s distal thigh and pushing it in the direction of the mother’s head. By leaning in the same direction, the clinician can use their own weight to facilitate hip hyperflexion. A common mistake is holding the mother’s knee and pushing at the foot. The resulting force is reduced and the mother may experience discomfort at the ankle joint (Fig a and b).9

Importantly, this study showed an increasing trend in the utilisation of posterior arm delivery during 2010-2019, which is consistent with the updated practice guidelines from the RCOG and the American College of Obstetricians and Gynecologists during the same period.1 2 The success rate of posterior arm delivery also substantially increased (from 52.9% to 92.3%) [Table 4] and humeral fracture (a complication associated with posterior arm delivery) did not occur in any case during 2010-2019. The improvement in the success and safety may be related to the enhanced training through the SOPHIE course, which emphasises that the clinician should use the correct hand (ie, right hand for a fetus facing the mother’s left side, and vice versa); this hand should be inserted into the vagina with sufficient depth to reach the fetal posterior forearm, and extraction should be conducted by grasping the forearm (rather than the elbow or upper arm). The direction of the extraction should be outward and upward to generate a rotational effect on the shoulders (Fig c and d).9 10

In addition to enhanced clinician skills, improved management of shoulder dystocia may have resulted from a decline in birth weight from 2000-2009 to 2010-2019, although the difference was not statistically significant. Improvements in the success rates of the McRoberts’ manoeuvre and posterior arm delivery led to improved management of shoulder dystocia, represented by a shorter HBDI (Table 4). Among cases of shoulder dystocia, the proportion of babies with an HBDI of ≤2 minutes increased from 59.0% during 2000-2009 to 79.6% during 2010-2019; conversely, the proportion of such babies with HBDI ≥5 minutes decreased from 9.7% to 5.6% (Table 4). Along with the improvement in HBDI, fewer babies had a low Apgar score. However, such improvements did not lead to significant reductions in the incidences of severe fetal acidosis and hypoxic-ischaemic encephalopathy, possibly because of the small sample size. The management of prenatally missed macrosomia with severe shoulder dystocia remains a substantial challenge. Our group recently reported a severe case of shoulder dystocia in which the posterior shoulder was also lodged in the middle of the pelvic cavity. The situation was resolved by our modified posterior axillary sling traction technique, which involved using a ribbon gauze to form a sling, in combination with long and slim right-angle forceps to facilitate sling placement.23 In our report as well as in our recent review, we emphasised that the sling primarily functions by facilitating shoulder rotation to enable delivery through the wider diagonal diameter of the outlet of the birth canal.10

The strengths of this study include its analysis of a relatively large number of women over two decades, the use of complete and audited outcomes data, and the uniform management of pregnancy complications in accordance with standard guidelines and department protocols.12 However, because this was a retrospective study, causal factors underlying the findings could not be fully elucidated; possible confounding factors included changes in the management of pregnancy complications, management of babies with macrosomia, and management of shoulder dystocia, as well as changes in the levels of skills and experience among clinical personnel during the study period.

More proactive management of macrosomic pregnancies led to decreases in the overall proportion of babies with macrosomia and incidence of shoulder dystocia from 2000-2009 to 2010-2019. Improvements in shoulder dystocia relief skills were demonstrated by increases in the success rates of manoeuvres (eg, the McRoberts’ manoeuvre and posterior arm delivery), as well as decreases in HBDI and Apgar scores.

All authors contributed to the concept or design of the study, acquisition of the data, analysis or interpretation of the data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content. All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

The corresponding author is the programme director for the Safe Obstetric Practice for High risk and Emergency (SOPHIE) course. Other authors have disclosed no conflicts of interest.

The authors thank the creators of the SOPHIE course (https://www.obg.cuhk.edu.hk/training-education/sophie-course/) and Ms Catherine Chan at the Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong for contributing to the illustrations.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Ethical approval was obtained from the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (Ref No.: CRE 2017.442). Informed patient consent was waived by the Committee due to the retrospective nature of the study.

1. Royal College of Obstetricians and Gynaecologists. Shoulder Dystocia. Green-Top Guideline No. 42. 2nd Edition. London: Royal College of Obstetricians and Gynaecologists. 2012. Available from: https://www.rcog.org.uk/media/ewgpnmio/gtg_42.pdf. Accessed 14 Aug 2023.

2. Practice Bulletin No 178: Shoulder Dystocia [editorial]. Obstet Gynecol 2017;129:e123-33. Crossref

3. Leung TY, Chung TK. Severe chronic morbidity of childbirth. Best Pract Clin Res Obstet Gynaecol 2009;23:401-23. Crossref

4. Leung TY, Stuart O, Sahota DS, Suen SS, Lau TK, Lao TT. Head-to-body delivery interval and risk of fetal acidosis and hypoxic ischaemic encephalopathy in shoulder dystocia: a retrospective review. BJOG 2011;118:474-9. Crossref

5. Leung TY, Stuart O, Suen SS, Sahota DS, Lau TK, Lao TT. Comparison of perinatal outcomes of shoulder dystocia alleviated by different type and sequence of manoeuvres: a retrospective review. BJOG 2011;118:985-90. Crossref

6. Hoffman MK, Bailit JL, Branch DW, et al. A comparison of obstetric maneuvers for the acute management of shoulder dystocia. Obstet Gynecol 2011;117:1272-8. Crossref

7. American Academy of Family Physicians. Advanced Life Support in Obstetrics (ALSO®). Available from: https://www.aafp.org/cme/programs/also.html. Accessed 7 Jan 2022.

8. PROMPT Maternity Foundation. Practical Obstetric Multi-Professional Training. Available from: https://www.promptmaternity.org. Accessed 7 Jan 2022.

9. Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong & Prince of Wales Hospital. SOPHIE Course–Safe Obstetric Practice for High risk and Emergency Course. Available from: https://www.obg.cuhk.edu.hk/training-education/sophie-course/. Accessed 14 Aug 2023.

10. Lau SL, Sin WT, Wong L, Lee NM, Hui SY, Leung TY. A critical evaluation of the external and internal maneuvers for resolution of shoulder dystocia. Am J Obstet Gynecol. In press. Crossref

11. Boulvain M, Senat MV, Perrotin F, et al. Induction of labour versus expectant management for large-for-date fetuses: a randomised controlled trial. Lancet 2015;385:2600-5. Crossref

12. Wong ST, Tse WT, Lau SL, Sahota DS, Leung TY. Stillbirth rate in singleton pregnancies: a 20-year retrospective study from a public obstetric unit in Hong Kong. Hong Kong Med J 2022;28:285-93. Crossref

13. Spong CY, Beall M, Rodrigues D, Ross MG. An objective definition of shoulder dystocia: prolonged head-to-body delivery intervals and/or the use of ancillary obstetric maneuvers. Obstet Gynecol 1995;86:433-6. Crossref

14. Leung TY, Lok IH, Tam WH, Leung TN, Lau TK. Deterioration in cord blood gas status during the second stage of labour is more rapid in the second twin than in the first twin. BJOG 2004;111:546-9. Crossref

15. Leung TY, Tam WH, Leung TN, Lok IH, Lau TK. Effect of twin-to-twin delivery interval on umbilical cord blood gas in the second twins. BJOG 2002;109:63-7. Crossref

16. WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004;363:157-63. Crossref

17. Leung TY, Leung TN, Sahota DS, et al. Trends in maternal obesity and associated risks of adverse pregnancy outcomes in a population of Chinese women. BJOG 2008;115:1529-37. Crossref

18. Cheng YK, Lao TT, Sahota DS, Leung VK, Leung TY. Use of birth weight threshold for macrosomia to identify fetuses at risk of shoulder dystocia among Chinese populations. Int J Gynaecol Obstet 2013;120:249-53. Crossref

19. McFarland MB, Langer O, Piper JM, Berkus MD. Perinatal outcome and the type and number of maneuvers in shoulder dystocia. Int J Gynaecol Obstet 1996;55:219-24. Crossref

20. Gherman RB, Goodwin TM, Souter I, Neumann K, Ouzounian JG, Paul RH. The McRoberts’ maneuver for the alleviation of shoulder dystocia: how successful is it? Am J Obstet Gynecol 1997;176:656-61. Crossref

21. Spain JE, Frey HA, Tuuli MG, Colvin R, Macones GA, Cahill AG. Neonatal morbidity associated with shoulder dystocia maneuvers. Am J Obstet Gynecol 2015;212:353.e1-5. Crossref

22. Lok ZL, Cheng YK, Leung TY. Predictive factors for the success of McRoberts’ manoeuvre and suprapubic pressure in relieving shoulder dystocia: a cross-sectional study. BMC Pregnancy Childbirth 2016;16:334. Crossref

23. Kwan AH, Hui AS, Lee JH, Leung TY. Intrauterine fetal death followed by shoulder dystocia and birth by modified posterior axillary sling method: a case report. BMC Pregnancy Childbirth 2021;21:672. Crossref

Primary ocular melanoma from melanocytes within the eye constitutes uveal melanoma (UM), whereas ocular melanoma from melanocytes on the globe surface is considered conjunctival melanoma (CM). Although both UM and CM develop from similar neural crest lineages and are classified as ocular melanomas, they have distinct clinical behaviours, management, prognosis, cancer staging features, and molecular characteristics. Overall, 85% of melanomas in ocular regions arise from the uvea (iris, choroidal, and ciliary body), 5% arise from the conjunctiva, and 10% occur in other sites.1 The most common site of UM is the choroid, which is involved in 90% of cases.2

For most cases of UM, the primary treatment is enucleation. Alternative options include plaque brachytherapy and proton beam radiation; these options are currently unavailable in Hong Kong and similar Asian countries (eg, Singapore). Systemic metastases of UM most commonly affect the liver, followed by lung and bone. Conjunctival melanoma is usually managed by complete excisional biopsy with wide surgical margins, using a ‘no-touch’ technique. Adjuvant therapies such as cryotherapy or topical mitomycin C are offered, followed by reconstruction with an amniotic membrane graft. Orbital exenteration may be necessary for advanced tumours where local resection is not feasible. Metastatic disease, often to regional lymph nodes and the brain, occurs in 20% to 30% of patients with CM.3

Melanomas are generally rare tumours and their incidences are particularly low in Asian populations. The annual incidence of UM in Caucasian populations is 5.1 cases per million,2 whereas the respective incidences in Japanese4 and Korean5 populations are 0.2 and 0.6 cases per million. The incidences of CM are similar: 0.2 to 0.5 cases per million in Caucasian populations3 and 0.15 cases per million in Asian populations.6 Because of this rarity among Asian populations, very few studies have focused on Chinese patients. Epidemiological studies have largely involved Caucasian populations. Clinical characteristics may differ in Asian populations; for example, Asian patients are initially diagnosed with UM5 and CM6 at younger ages.

Known risk factors for UM include fair skin, light-coloured eyes, congenital ocular melanocytosis, ocular melanocytoma, and BAP1 tumour predisposition syndrome.7 In contrast, risk factors for CM include increased conjunctival pigmentation and a history of primary acquired melanosis. Predictors of recurrence or new tumour formation after CM treatment are older age, a history of prior conjunctival surgery, and advanced T subclassification in the American Joint Committee on Cancer (AJCC) staging system.8

This study explored the disease course and outcomes of UM and CM in Chinese patients without nodal metastasis on presentation.

This retrospective study included patients who received treatment for UM or CM in a single tertiary ophthalmic centre in Hong Kong between January 1994 and December 2019. Inclusion criteria included Chinese ethnicity and imaging-confirmed lack of tumour dissemination at diagnosis. Exclusion criteria were <6 months of follow-up, insufficient available information, loss to follow-up, or presence of any precursor lesions (eg, atypical primary acquired melanosis). The following data were recorded: patient demographics, tumour laterality, tumour characteristics (eg, presentation and staging according to the AJCC Cancer Staging Manual [8th Edition]9 10), investigations performed, treatment regimen, and final outcomes. The study adhered to the principles outlined in the Declaration of Helsinki.

Provisional clinical diagnoses of UM and CM were made based on each patient’s medical history, primary acquired melanosis status, and clinical presentation. Final diagnoses of UM and CM were confirmed by excisional biopsy or analysis of a specimen collected during definitive surgical treatment. Unless refused by the patient, positron emission tomography–computed tomography (PET-CT) scans of UM and CM were performed after 2001 (when such scans became commercially available). Magnetic resonance imaging scans of the brain and orbit, as well as fundus fluorescein angiography and indocyanine green angiography, were conducted to investigate suspected UM. All patients with pathologically confirmed UM or CM underwent definitive surgical treatment. The specific surgical treatment was selected according to melanoma location and size, depth of invasion, systemic metastasis status, and the patient’s physical condition.

SPSS software (Windows version 20; IBM Corp, Armonk [NY], US) was utilised for statistical analysis. Differences between patient groups were calculated using the Chi squared and Mann-Whitney U tests. P values <0.05 were considered statistically significant. Kaplan-Meier survival analysis was performed to estimate disease-free survival and overall survival in patients who had received definitive treatment. Continuous data were reported as mean ± standard deviation.

In this 25-year retrospective study, there were 13 and 11 patients with pathologically confirmed UM and CM, respectively. Two other patients were excluded: both displayed ocular symptoms (upper eyelid mass and bloody discharge) but were subsequently diagnosed with ocular metastases of primary nasal melanoma. The mean follow-up durations for UM and CM were 67 ± 53.2 and 74 ± 83.3 months, respectively. There were more men among patients with UM than among those with CM (P=0.042). The incidence of pre-existing conjunctival nevus was higher among patients with CM than among those with UM (P=0.049). There were no other significant differences in terms of age at diagnosis, tumour laterality, duration of symptoms, history of ocular nevi, or pathologically determined lesion diameter (Table 1).

Twenty-three of the 24 included patients had visual symptoms on presentation. Among patients with UM, 84.6% (n=11) reported blurring of vision and 7.7% (n=1) presented with photopsia. In one male patient (7.7%), the tumour was discovered during a routine ophthalmological examination. Patients with CM had more diverse visual symptoms: 63.6% (n=7) exhibited conjunctival pigmentation, 18.2% (n=2) had either an upper or lower eyelid mass, 9.1% (n=1) displayed bloody ocular discharge, and 9.1% (n=1) had experienced bleeding from the mass.

Retrospective melanoma staging of UM9 and CM10 was conducted in accordance with the AJCC Cancer Staging Manual (8th Edition). Uveal melanoma stages varied from T1a to T4a and from stage I to IIIa. Most UM cases were clinical stage II: 38.5% (n=5), 38.5% (n=5), 7.7% (n=1), and 15.3% (n=2) of patients with UM had clinical stage I, IIA, IIB, and IIIA tumours, respectively. Pathological staging of the tumours revealed that 46.2% (n=6) were spindle cell type, 38.5% (n=5) were epithelioid cell type, and the remaining 15.3% (n=2) were mixed cell type (ie, >10% epithelioid cells and <90% spindle cells). Pathological staging tended to be equivalent to or higher than the initial clinical staging: two patients with pathological stage IIB disease were initially diagnosed with clinical stage I disease, and three patients with pathological stage IIB disease were initially diagnosed with clinical stage IIA disease. For the remaining eight patients, the clinical and pathological stages were identical. Overall, 23.1% (n=3), 15.4% (n=2), 46.2% (n=6), and 15.3% (n=2) of patients with UM had pathological stage I, IIA, IIB and III tumours, respectively.

At diagnosis of CM, the clinical T (cT) stage was cT2 in 72.7% (n=8) of patients and cT3 in 27.3% (n=3) of patients. There were no cT1 or cT4 tumours in our cohort. The pathological T (pT) stage was pT2 in 54.5% (n=6) of patients and pT3 in 45.5% (n=5) of patients. No patients were diagnosed with pT1 or pT4 disease. Unlike the approach or UM, the AJCC staging system for CM does not include guidance regarding overall stage; there is only clinical and pathological staging for the T (tumour) component. One patient had a lower clinical stage (T2b) than pathological stage (T3b); all other patients had identical clinical and pathological stages.

In terms of disease management, biopsies were more frequently performed before definitive treatment in patients with CM than in those with UM (P=0.003). Enucleation was performed in 92.3% (n=12) of patients with UM, whereas orbital exenteration was performed in 81.8% (n=9) of patients with CM. The interventions and treatments performed are shown in Table 2.

After definitive treatment, all patients initially attended weekly follow-up visits; they gradually transitioned to follow-up at 6-month intervals. Clinical examinations were performed for any surgical complications or disease recurrence. In cases of suspected disease recurrence or metastasis, contrast CT scans of the brain and orbit were performed. No patients with UM or CM had short-term or long-term wound complications. There were no instances of local recurrence during follow-up.

One patient with UM (7.7%) had metastasis to bone, lung, and breast tissue, as determined by high-resolution CT at 105 months of follow-up. The patient died 117 months after initial diagnosis. Two patients (15.3%) died of causes unrelated to melanoma, such as hepatocellular carcinoma or squamous cell carcinoma of the lung.

Patients with CM had worse outcomes than those with had UM, in terms of systemic metastasis (P=0.031) and tumour-related mortality rates (P=0.006). Overall, 45.5% (n=4) of patients with CM developed systemic metastasis during follow-up; the mean time from definitive treatment to systemic metastasis was 55.8 ± 66 months (range, 14-168). Cases were detected when patients were symptomatic and admitted to an acute care hospital for whole-body PET-CT or brain magnetic resonance imaging. Lymph node metastasis in two patients and liver metastasis in one patient were confirmed by fine needle aspiration cytology. Brain metastasis was identified by brain CT in two patients, one of whom had simultaneous bone and lung metastases confirmed by PET-CT. Overall, 18.2% (n=2) of patients died of causes unrelated to melanoma. The mean time from definitive treatment to death was 39 ± 26 months (range, 1-64). Patients with CM had shorter median disease-free survival (P=0.004) and overall survival (P=0.006). Detailed results are presented in Table 3.

At 2 years after definitive treatment, the probabilities of disease-free survival were approximately 0.55 for CM and 1.0 for UM. At 10 years, these probabilities decreased to 0.4 for CM and 0.7 for UM. Overall survival was 100% among patients with UM at 10 years after definitive treatment. In contrast, patients with CM displayed a progressive decrease in overall survival, reaching 35% at 5 years after definitive treatment. Survival curves are depicted in Figures 1 and 2.

To our knowledge, this is the first retrospective study of disease course and outcomes among patients with UM and CM in southern China.

In the present study, the mean ages at diagnosis of UM and CM were 59 and 57 years, respectively. These findings are similar to the ages at diagnosis of UM in Taiwan (55 years)11 and the US (58 years),12 but slightly older than the ages in Singapore (52 years)13 and South Korea (53 years).14 The findings are also similar to the ages at diagnosis of CM in the US and mainland China (61 years15 and 54 years,16 respectively). It is clear that UM and CM both affect patients in their late 50s to early 60s, regardless of ethnicity.

Uveal melanoma primarily affected men in the present study, consistent with findings in Australian17 and European18 Caucasian populations. No previous studies have identified oestrogen receptors in normal uveal tissue or UM tumours,19 and there is no evidence that oral contraceptives or postmenopausal oestrogens participate in UM aetiology.20 Because female hormones do not have protective or exacerbating roles in UM, there is speculation that testosterone receptors are present on UM tumours, leading to a higher incidence in men.21 The present study revealed that CM primarily affected women, but previous studies have not found a relationship between sex and CM incidence.22 23 In contrast to our results, a study in mainland China showed that CM primarily affected men.15 Further studies are needed regarding the relationship between sex and CM.

Studies in Caucasian populations have demonstrated that the incidence of UM is much higher than that of CM, with a ratio of approximately 3 to 1.21 24 In the present study, CM constituted a larger proportion of ocular melanoma cases than previously identified in Caucasian populations, with results similar to epidemiological findings from Korea.5 Asian populations may have a lower risk of UM, but further research is warranted to confirm this speculation.

Zloto et al21 described an intriguing phenomenon whereby men were less likely than women to report symptoms of UM. Among our patients with UM, nearly all reported symptoms; one male patient did not report symptoms and had a tumour identified during a routine examination. Although this finding was not statistically significant, it is consistent with the previous results in a Caucasian population.2

In the present study, more patients with CM had pre-existing conjunctival pigmented lesions, compared with those who had UM. Somatic mutations in the BRAF gene frequently occur in human melanomas, including CM; such mutations are strongly associated with ultraviolent light exposure.25 In subtropical regions such as Hong Kong where there is abundant sunlight, analyses of underlying BRAF gene mutations could be insightful.

Among 10 ophthalmology centres on four continents (North and South America, Europe, and Asia), the proportions of UM stages I, IIA, IIB, and IIIA were 32%, 34%, 22.1%, and 8.8%, respectively10; the present study demonstrated a comparatively greater proportion of cases with stage IIB or higher.

Comparative analysis of CM data revealed similar results: patients in the present study had higher clinical stages relative to patients in the US, where three-quarters and more than half of the patients had clinical and pathological stage I disease. This discrepancy could be explained by the rare nature of CM, particularly in Asian populations, leading to lower disease awareness and delayed referral to an ophthalmologist.

Primary enucleation was an effective treatment for patients with UM in the present study; only one patient (7.7%) had systemic metastasis. In that patient, although no systemic metastasis was detected on PET-CT at diagnosis, the maximum standardised uptake value of the UM tumour was particularly high (9.4); UM tumours in other patients had maximum standardised uptake values of 0 to 3.9. Considering the exceptionally high metabolic rate in the UM tumour of the patient with systemic metastasis, microscopic metastasis may have been present before enucleation.

Despite radical treatment with wide excisional biopsy or primary orbital exenteration, systemic metastasis occurred in nearly half of the patients with CM. Regular imaging surveillance by PET-CT may be beneficial for patients with CM after orbital exenteration.

Among patients with UM, we observed a much lower rate of systemic metastasis than reported in Singapore (7.7% vs 45.5%26). This difference could be attributed to the performance of early radical treatment (ie, enucleation) before detection of systemic metastasis. Notably, the disease-free survival rate was better in our cohort of patients with UM than in another study of Chinese patients with UM (5- and 10-year disease-free survival rates of 80% and 70%, respectively)27 and better than in a study of Singaporean patients with UM (5-year disease-free survival rate of 56.8%).13 For comparison, in Caucasian populations, the 5-year and 10-year disease-free survival rates were 81.6%17 and 50%,28 respectively.

The rates of systemic metastasis and tumour-related mortality are consistently higher in patients with CM than in those with UM. In the present study, tumour-related mortality at 5 years was similar to the rate in Chinese patients (30.5%)16 but higher than that among patients in the US (7%).29 In five large studies (n=734 cases overall) of CM after surgical resection with tumour-free margins, the 5-year overall survival rates ranged from 74% to 86%.30 31 32 33 34 A recent Singaporean study revealed a slightly lower 5-year overall survival rate (68.8%).13 In the present study, the 5-year overall survival rate was substantially lower than the rates in other countries. Shields et al29 reported that a pathologically confirmed positive tumour margin and the absence of limbal involvement were risk factors for CM metastasis. Although all patients with CM in our cohort had pathologically confirmed clear margins, most patients with metastasis (n=6) had palpebral CM (83.3%, n=5) in which the melanoma did not reach the limbus. Thus, tumour location and ethnicity may explain the poor overall survival among patients with CM in the present study.

The higher risk of metastasis and lower rates of 5- and 10-year overall survival in CM, compared with UM, could be attributed to multiple factors. We did not find significant differences between UM and CM in terms of tumour stage or delays in diagnosis/treatment. We suspect that the differences between tumours are related to the nature of the disease, the primary mode of metastasis (UM spreads through the vasculature, whereas CM spreads through the lymphatic system), and genetic alterations (UM is associated with chromosomal abnormalities and CM is associated with mutations in specific genes). In terms of monitoring UM recurrence, the use of single-cell technologies to identify circulating tumour cells has implications for clinical stratification, particularly in cases of UM where specific genetic mutations have been identified.35 Because circulating tumour cell tests have received US Food and Drug Administration’s approval for clinical use in the management of various tumours (eg, metastatic breast and prostate cancers), they may be utilised in future efforts to detect circulating UM tumour cells.

In our centre, sentinel lymph node biopsy (SLNB) is not performed as a component of CM management. Mor et al36 recommend SLNB for patients with CM because false-negative findings are rare and 5-year survival can reach 79%. Therefore, early diagnosis of CM, including SLNB in cases with poor prognosticating factors (lack of limbal involvement and positive biopsy margin), and radical neck dissection as appropriate (with support from head and neck surgeons) should be considered before systemic treatment is offered.

This retrospective study included a low number of patients. Additionally, genetic testing was not performed on tissue samples from patients with pathologically confirmed tumours. Chromosomal abnormalities (monosomy 3, gain of chromosome 8q, and monosomy 3 combined with loss of 1p36) have been associated with decreased survival in UM,37 whereas mutations in the BRAF, RAS, cKIT, and NF1 genes have been associated with CM38; thus, survival could be more closely related to specific genetic features, and it may be inappropriate to consider these tumours as single entities.

Because UM and CM are rare conditions, they represent challenges for primary physicians (ie, timely referral) and ophthalmologists (ie, appropriate treatment and adequate long-term follow-up). Currently, there is limited information regarding the roles of newer targeted therapies for UM and CM, compared with the application of such therapies to cutaneous melanoma. Among patients with CM, long-term mortality remains high despite definitive radical treatment. This study explored the disease course and outcomes in Hong Kong Chinese patients, then compared the findings with data from patients in other countries. For patients with UM and CM, we recommend long-term follow-up with close monitoring, a detailed medical history, holistic assessment involving cervical and head lymph node palpation, and ophthalmological examination. Collaborations with oncologists to provide regular systemic evaluation during long-term followup, with the goal of early detection for distant metastases, are also important. Chest X-ray, brain magnetic resonance imaging, and cytology with SLNB should be performed regularly (annually if possible) to improve survival, particularly in patients with CM.

Concept or design: HKL Yuen.
Acquisition of data: JYY Chan.
Analysis or interpretation of data: SC Lam, JYY Chan.
Drafting of the manuscript: JYY Chan.
Critical revision of the manuscript for important intellectual content: SC Lam, HKL Yuen.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The study protocol was approved by the Research Ethics Committee (Kowloon Central/Kowloon East) of Hospital Authority, Hong Kong (Ref No.: KC/KE-21-0129/ER-1). Patients were treated in accordance with the tenets of the Declaration of Helsinki, provided written informed consent for all treatments and procedures, and consented to publication of this report.

1. Chang AE, Karnell LH, Menck HR. The National Cancer Data Base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer 1998;83:1664-78. Crossref

2. Kaliki S, Shields CL. Uveal melanoma: relatively rare but deadly cancer. Eye (Lond) 2017;31:241-57. Crossref

3. Vora GK, Demirci H, Marr B, Mruthyunjaya P. Advances in the management of conjunctival melanoma. Surv Ophthalmol 2017;62:26-42. Crossref

4. Stang A, Parkin DM, Ferlay J, Jöckel KH. International uveal melanoma incidence trends in view of a decreasing proportion of morphological verification. Int J Cancer 2005;114:114-23. Crossref

5. Park SJ, Oh CM, Kim BW, Woo SJ, Cho H, Park KH. Nationwide incidence of ocular melanoma in South Korea by using the national cancer registry database (1999-2011). Invest Ophthalmol Vis Sci 2015;56:4719-24. Crossref

6. Hu DN, Yu G, McCormick SA, Finger PT. Population-based incidence of conjunctival melanoma in various races and ethnic groups and comparison with other melanomas. Am J Ophthalmol 2008;145:418-23. Crossref

7. Jager MJ, Shields CL, Cebulla CM, et al. Uveal melanoma. Nat Rev Dis Primers 2020;6:24. Crossref

8. Vaidya S, Dalvin LA, Yaghy A, et al. Conjunctival melanoma: risk factors for recurrent or new tumor in 540 patients at a single ocular oncology center. Eur J Ophthalmol 2021;31:2675-85. Crossref

9. Baron ED, Di Nicola M, Shields CL. Updated AJCC classification for posterior uveal melanoma. Available from: https://retinatoday.com/pdfs/0518RT_Oncology.pdf. Accessed 1 Mar 2021.

10. Jain P, Finger PT, Damato B, et al. Multicenter, international assessment of the Eighth Edition of the American Joint Committee on Cancer Cancer Staging Manual for conjunctival melanoma. JAMA Ophthalmol 2019;137:905-11. Crossref

11. Ma ST, Hsieh YT, Wei YH, Liao SL. A 45-year experience of uveal melanoma in Taiwan: verification of American Joint Committee on Cancer staging system and prognostic factors. J Formos Med Assoc 2021;120:1361-8. Crossref

12. Shields CL, Furuta M, Thangappan A, et al. Metastasis of uveal melanoma millimeter-by-millimeter in 8033 consecutive eyes. Arch Ophthalmol 2009;127:989-98. Crossref

13. Tan LL, Hong J, Goh WL, et al. Clinical features and survival outcomes of ocular melanoma in a multi-ethnic Asian cohort. Sci Rep 2020;10:16367. Crossref

14. Lee CS, Lee J, Choi JJ, et al. Cytogenetics and prognosis for uveal melanoma in Korean patients. Acta Ophthalmol 2011;89:e310-4. Crossref

15. Yousef YA, Finger PT. Predictive value of the seventh edition American Joint Committee on Cancer staging system for conjunctival melanoma. Arch Ophthalmol 2012;130:599-606. Crossref

16. Zhou C, Wang Y, Jia R, Fan X. Conjunctival melanoma in Chinese patients: local recurrence, metastasis, mortality, and comparisons with Caucasian patients. Invest Ophthalmol Vis Sci 2017;58:5452-9. Crossref

17. Vajdic CM, Kricker A, Giblin M, et al. Incidence of ocular melanoma in Australia from 1990 to 1998. Int J Cancer 2003;105:117-22. Crossref

18. Singh AD, Turell ME, Topham AK. Uveal melanoma: trends in incidence, treatment, and survival. Ophthalmology 2011;118:1881-5. Crossref

19. Mäkitie T, Tarkkanen A, Kivelä T. Comparative immunohistochemical oestrogen receptor analysis in primary and metastatic uveal melanoma. Graefes Arch Clin Exp Ophthalmol 1998;236:415-9. Crossref

20. Holly EA, Aston DA, Ahn DK, Kristiansen JJ, Char DH. Uveal melanoma, hormonal and reproductive factors in women. Cancer Res 1991;51:1370-2.

21. Zloto O, Pe’er J, Frenkel S. Gender differences in clinical presentation and prognosis of uveal melanoma. Invest Ophthalmol Vis Sci 2013;54:652-6. Crossref

22. McLaughlin CC, Wu XC, Jemal A, Martin HJ, Roche LM, Chen VW. Incidence of noncutaneous melanomas in the U.S. Cancer 2005;103:1000-7. Crossref

23. Wong JR, Nanji AA, Galor A, Karp CL. Management of conjunctival malignant melanoma: a review and update. Expert Rev Ophthalmol 2014;9:185-204. Crossref

24. Hu DN, Yu GP, McCormick SA, Schneider S, Finger PT. Population-based incidence of uveal melanoma in various races and ethnic groups. Am J Ophthalmol 2005;140:612-7. Crossref

25. Besaratinia A, Pfeifer GP. Sunlight ultraviolet irradiation and BRAF V600 mutagenesis in human melanoma. Hum Mutat 2008;29:983-91. Crossref

26. Wong W, Sundar G, Chee C, Zhao PS, Rajagopalan R, Gopal L. Clinical spectrum, treatment and outcomes of uveal melanoma in a tertiary centre. Singapore Med J 2019;60:474-8. Crossref

27. Yue H, Qian J, Yuan Y, et al. Clinicopathological characteristics and prognosis for survival after enucleation of uveal melanoma in Chinese patients: long-term follow-up. Curr Eye Res 2017;42:759-65. Crossref

28. Virgili G, Gatta G, Ciccolallo L, et al. Survival in patients with uveal melanoma in Europe. Arch Ophthalmol 2008;126:1413-8. Crossref

29. Shields CL, Shields JA, Gündüz K, et al. Conjunctival melanoma: risk factors for recurrence, exenteration, metastasis, and death in 150 consecutive patients. Arch Ophthalmol 2000;118:1497-507. Crossref

30. Tuomaala S, Eskelin S, Tarkkanen A, Kivelä T. Population-based assessment of clinical characteristics predicting outcome of conjunctival melanoma in whites. Invest Ophthalmol Vis Sci 2002;43:3399-408.

31. Missotten GS, Keijser S, De Keizer RJ, De Wolff-Rouendaal D. Conjunctival melanoma in the Netherlands: a nationwide study. Invest Ophthalmol Vis Sci 2005;46:75-82. Crossref

32. Shields CL, Kaliki S, Al-Dahmash SA, Lally SE, Shields JA. American Joint Committee on Cancer (AJCC) clinical classification predicts conjunctival melanoma outcomes. Ophthal Plast Reconstr Surg 2012;28:313-23. Crossref

33. Esmaeli B, Wang X, Youssef A, Gershenwald JE. Patterns of regional and distant metastasis in patients with conjunctival melanoma: experience at a cancer center over four decades. Ophthalmology 2001;108:2101-5. Crossref

34. Werschnik C, Lommatzsch PK. Long-term follow-up of patients with conjunctival melanoma. Am J Clin Oncol 2002;25:248-55. Crossref

35. Wang MM, Chen C, Lynn MN, et al. Applying single-cell technology in uveal melanomas: current trends and perspectives for improving uveal melanoma metastasis surveillance and tumor profiling. Front Mol Biosci 2021;7:611584. Crossref

36. Mor JM, Rokohl AC, Koch KR, Heindl LM. Sentinel lymph node biopsy in the management of conjunctival melanoma: current insights. Clin Ophthalmol 2019;13:1297-302. Crossref

37. van den Bosch T, Kilic E, Paridaens D, de Klein A. Genetics of uveal melanoma and cutaneous melanoma: two of a kind? Dermatol Res Pract 2010;2010:360136. Crossref

38. Gkiala A, Palioura S. Conjunctival melanoma: update on genetics, epigenetics and targeted molecular and immune-based therapies. Clin Ophthalmol 2020;14:3137-52. Crossref

Paediatric intensive care units (PICUs) are highly specialised workplaces that support children with critical illnesses and their caregivers. Advances in paediatric critical care have significantly improved survival among critically ill children, although this improvement has also led to higher rates of morbidity, more disabilities, and longer hospital stays.1 2 3 4 5 These changes have resulted in potentially conflicting views regarding expectations and treatment goals among healthcare workers and patients’ families, increasing the incidence of moral distress among healthcare workers.6

Moral distress is a term that refers to experiences of frustration and failure arising from healthcare workers’ attempts to fulfil their moral obligations to patients, families, and the public.7 8 In an intensive care setting, healthcare workers frequently encounter ethical issues. Moral distress arises when a healthcare worker has determined the right course of action but cannot follow it because of internal or external constraints (eg, limited resources, institutional policies, or family preferences).9 Moral distress has been identified among healthcare workers in both adult ICUs and PICUs.10 11 It is associated with greater experience and lower staff retention.12

Depression and stress-related symptoms are common in healthcare workers, particularly among ICU staff.13 14 Studies have shown that these symptoms can ultimately impair patient care quality.15 16 Thus far, most literature regarding moral distress has been published in Western countries; the concept of moral distress is not well-known outside of the Western world.17 To our knowledge, there have been few analyses of moral distress and psychological status among healthcare workers in non-Western PICUs. Factors that can influence the level and type of moral distress include cultural backgrounds; beliefs of the patient, their family, and the clinical team; and differences among healthcare systems. Hong Kong is a multicultural city influenced by both Eastern and Western cultures; challenges in this setting may be unique. This study assessed moral distress prevalence and psychological status among PICU healthcare workers in Hong Kong.

This prospective single-centre cross-sectional study was conducted from June to July 2020 in the six-bed tertiary PICU of the Hong Kong Children’s Hospital (HKCH), which began operation at the end of March 2019. The HKCH is the only dedicated paediatric oncology centre in the region, and most PICU admissions (54%) during the study period involved patients with cancer.

Study participants were healthcare workers involved in direct clinical care within the HKCH PICU, including doctors, nurses, and allied health professionals (ie, physiotherapists, occupational therapists, speech therapists, pharmacists, and dietitians). Healthcare workers were excluded if they had <3 months of critical care experience in the PICU or were temporarily on leave from the PICU during the study period. The survey was distributed to all eligible healthcare workers in the HKCH PICU during working hours within the study period.

The survey included two validated instruments (Revised Moral Distress Scale [MDS-R] Paediatric Version and Depression Anxiety and Stress Scale–21 items [DASS-21]) to measure levels of moral stress, depression, anxiety, and stress in all participants.18 19 The participants’ demographic details were also collected. The survey explored job-quitting intentions related to moral distress or other reasons. It was piloted with two HKCH PICU staff members; questions were refined based on feedback from them. The final survey was paper-based. An email was sent to all participants before study commencement with information regarding the aim and details of the study. The survey was distributed by hand, and all copies were collected in a sealed box after completion. To ensure anonymity, the survey did not contain any identifiers.

Moral distress, the main outcome of the study, was measured using the validated paediatric version of the MDS-R (online supplementary Appendix 1).18 It consists of 21 items describing predetermined potentially morally distressing situations. There are five predetermined categories of situations: end-of-life care and quality of life, poor communication, staffing and material resources, hierarchies of decision making, and witnessing unethical behaviour. Each item on the MDS-R is scored according to the frequency and intensity that a healthcare worker experienced, using a Likert scale that ranges from 0 to 4. If a specific situation has never been experienced, participants are asked to indicate how disturbing the situation would be if they encountered it in their workplace. The frequency and intensity scores are then multiplied to produce an overall score for each item. The total moral distress score is the sum of the 21 overall scores for each item, ranging from 0 to 336. The English version of this instrument was used.

Psychological status was assessed using the DASS-21 (online supplementary Appendix 2).19 It is a set of three self-reporting subscales that measure participants’ emotional states: depression, anxiety, and stress. Each scale contains seven items for each emotional state. Each item is scored on a four-point Likert scale ranging from 0 (‘Did not apply to me at all’) to 3 (‘Applied to me very much or most of the time’). The total score for each emotional state is the sum of the subscale scores multiplied by 2. Depression, anxiety, or stress was considered present if the relevant scores exceeded the normal cut-off. The emotional state was categorised as mild, moderate, severe, or extremely severe, based on published cut-offs. The English and Chinese versions of this instrument were used; both language versions have been validated.19 20

Outcome measures were demographic data and the levels of moral distress, depression, anxiety, and stress. Data were expressed using median (interquartile range [IQR]) for continuous variables and count (percentage) for categorical variables. Results of the MDS-R and DASS-21 were compared among doctors, nurses, and allied health professionals using the Chi squared test, Kruskal-Wallis test, or Cohen’s d. Correlations between participant variables and outcome measures were evaluated using Spearman’s rank correlation coefficient. P values <0.05 were considered statistically significant. Statistical analysis was performed using SPSS (Windows version 26.0; IBM Corp, Armonk [NY], United States).

In total, 46 of 56 healthcare workers in the PICU completed the survey; the response rate was 82%. On one survey, the moral distress section was incomplete; that survey was excluded from the analysis of moral distress.

Most participants were women (n=36, 78%) and were aged ≥30 years (n=35, 76%). More than half of the participants were nurses (n=26, 57%). Approximately half of the participants (n=24, 52%) had >5 years of PICU experience. Detailed participant characteristics are presented in Table 1.

The median MDS-R score was 71 (IQR=34-115). There was a significant difference in MDS-R score among the three professions (P<0.001). Doctors and nurses had significantly higher MDS-R scores, compared with allied health professionals (P<0.05). Nurses had the highest median MDS-R score (102, IQR=71-126), whereas allied health professionals had the lowest (20, IQR=6-39). There were no significant differences in MDS-R score according to sex, age, or duration of PICU experience (Table 1).

Among the 21 items on the MDS-R, the most morally distressing item was related to end-of-life care and quality of life: ‘Honour the family’s wishes to continue life support even though I believe it is not in the child’s best interest’. This item also scored highest in frequency and intensity among the 21 items. All three groups of health professionals ranked this item as the most morally distressing situation in the clinical setting. The second most morally distressing item was also related to end-of-life care and quality of life: ‘Initiate extensive life-saving actions when I think they only prolong death’. This item also consistently scored high in frequency and intensity (Table 2). Situations involving poor communication constituted the remaining three most morally distressing items in this study. The top five most morally distressing items, as well as the top five items with the highest frequency and intensity, are presented in Table 2.

A higher MDS-R moral distress score was associated with the intention to quit. Healthcare workers who intended to quit their jobs had significantly higher moral distress scores (P<0.05). A higher moral distress score was also associated with higher DASS-21 depression factor (r=0.445; P<0.05) and anxiety factor scores (r=0.417; P<0.05). Nurses who had worked for a greater number of years in the PICU also experienced higher moral distress (r=0.512; P<0.05). Twenty-eight percent of all participants and 35% of nurses reported they intended to quit their jobs because of moral distress.

The median depression, anxiety, and stress scores were 11 (IQR=0.5-18), 8 (IQR=3-145), and 30 (IQR=21-38), respectively; these scores corresponded to mild depression, mild anxiety, and severe stress. Among the three groups, nurses had the highest median anxiety (11, IQR=6-16) and stress scores (20, IQR=12-26) [Fig]; these scores corresponded to mild depression, moderate anxiety, and moderate stress. Participants with significantly higher depression and anxiety (both P<0.05) scores also intended to quit their jobs. There was no significant difference in stress score between participants who did and did not intend to quit their jobs (P=0.434).

In this study, various levels of moral distress were present in all three groups of PICU healthcare workers. There was a significant difference in MDS-R scores among the three professions, and nurses had the highest median MDS-R score. This finding is contrary to the results of previous PICU studies, which showed that moral distress did not differ among various healthcare workers.21 22 The literature suggests that nurses exhibit higher moral distress scores because they often have less autonomy concerning options in situations that involve moral dilemmas, and they are required to implement care plans with which they do not agree.23 24 25 26 Studies of PICU healthcare workers’ behaviours in ethical and morally distressing dilemmas have shown that 48% of PICU nurses reported needing to perform actions that violated their conscience. These results reflect the culture and hierarchies of power in the PICU.23 26 27 Moreover, nurses are the frontline workers who directly experience the impacts of clinical decisions on patients and their families.26 28 In newly established PICUs, decreased self-confidence or increased fear in a new working environment, combined with an uncertain ethical climate, unclear team dynamics, and less decision-making autonomy regarding care plans, can cause nurses to perceive less moral agency (ie, ability to act morally and change a situation).22 24 25 26 29 30 31 A reduced sense of moral agency can result in moral distress, which may be more apparent in newly established PICUs.29 31

We note that moral distress scores among nurses in the present study are among the highest in published studies of PICU healthcare workers (Table 3). In addition to the aforementioned lack of clarity in working environment and team dynamics, the diverse levels of experience among nurses might have also contributed to their high moral distress scores. In the present study, 54% of nurses had <3 years of PICU experience, whereas 39% of nurses had >10 years of PICU experience. These proportions of nurses with extensive and minimal experience were both larger than the proportions reported in previous PICU studies.32 33 The presence of such a large number of inexperienced junior nurses in the PICU may place additional stress on more experienced nurses. Indeed, survey items related to staffing (item 17 ‘Work with nurses or other care providers who are less competent than the child’s care requires’ and item 21 ‘Work with levels of care provider staffing that I consider unsafe’ in the MDS-R) were ranked by nurses as the seventh and eighth most morally distressing items; these rankings were higher than in other professions.

The PICU case mix might also contribute to moral distress. The majority of PICU admissions during the study period involved patients with cancer, who had considerably higher mortality rates; care for such patients frequently involved end-of-life and palliative care issues.34 35 In a study of nurses’ experiences while caring for dying children, Davies et al36 found that when nurses recognise a child’s death is inevitable, they often have to manage conflicting obligations: follow the doctor’s treatment orders and allow the child to die without unnecessary pain. These disparate treatment goals for critically ill children with terminal cancer can exacerbate moral distress.36 37 In a comparison of moral distress scores among various paediatric disciplines (eg, general care and surgical service), Trotochaud et al21 found that healthcare workers in haematology/oncology areas experienced the second highest amount of moral distress on the list, second to healthcare workers in PICUs. Moreover, the proportion of patients with cancer in our PICU is much higher than the proportions in previous PICU studies.38 39 Therefore, it is entirely understandable that moral distress in our PICU was particularly high among nurses.

The present study revealed a positive correlation between years of PICU experience and moral distress scores among nurses, consistent with previous results concerning healthcare workers in PICUs and adult ICUs.12 26 This correlation may be related to effective utilisation of clinical knowledge and experience, along with greater awareness concerning the impacts of potentially inappropriate treatment plans on patients.40 Conversely, a study by Larson et al26 revealed a negative correlation between moral distress scores and years of experience among doctors in the PICU. However, the present study showed no correlation between moral distress scores and years of experience among doctors. This finding might be attributed to the small number of doctors involved, which was insufficient to demonstrate an association.

Moral distress is often associated with the intention to quit a job.41 42 43 44 The results of the study were consistent with previous findings. Studies by Sannino et al11 and Trotochaud et al21 showed that 10.3% to 25% of PICU nurses intended to quit their jobs because of moral distress. The proportion of nurses in our study who intended to quit their job because of moral distress (34.6%) was higher than the proportions in previous PICU studies,11 21 which could be explained by their high moral distress scores. However, further studies are needed to determine the impact of moral distress alone on a healthcare worker’s intention to quit their job, compared with other possible distressing factors (eg, working hours and promotional opportunities) that can have a synergistic effect on the decision to quit.

To our knowledge, this is the first study of moral distress among healthcare workers in an East Asian PICU. The results of this study provide insights concerning the broader understanding of moral distress in newly established PICUs. The high response rate also suggests strong participation and indicates that the study sample is representative of healthcare workers in our PICU.

However, the results of this study should be interpreted with the following caveats. First, this was a single-centre study with a relatively small sample size, which limits the generalisability of the findings. The small sample size also hindered further evaluation of identifiable demographic factors, such as education level and whether participants had any children; another study indicated that such factors may be associated with moral distress.11 Moreover, the small sample size precluded subgroup analysis. Second, this study was susceptible to ‘survivorship’ bias because the sample did not include PICU staff who already quit their jobs, including some who quit because of moral distress. Third, considering the cross-sectional nature of this study, causal relationships among various factors could not be established. For example, although participants with higher depression and anxiety scores reported a stronger intention to quit their jobs, we could not determine whether these participants reported more psychological symptoms because of their intention to quit, or if their intention to quit led to more psychological symptoms. Larger multicentre studies are needed to further explore moral distress among healthcare workers in Hong Kong PICUs. As our unit expands to a 16-bed PICU and a five-bed high-dependency unit, a longitudinal study will also enhance the broader understanding of moral distress dynamics in a developing PICU, as well as the efficacies of various strategies to address moral distress.

Considering the results of this study, moral distress should be regarded as a key area for service improvement. The high levels of moral distress experienced by nurses, as well as the substantial moral distress in relation to end-of-life care, suggest that coping strategies should target nurses and focus on end-of-life education. These coping strategies are urgently needed to improve staff retention and quality of care; they can be implemented at the individual, organisational, and administrative levels.20

At the individual level, ethics education is essential for improvements in coping capacity and sense of moral agency, which can reduce the levels of moral distress.22 45 Education can be provided through interactive workshops or self-guided programmes.41 Prentice et al42 suggested that education should focus on improving knowledge regarding patient outcomes, the degree of uncertainty in specific situations, and appropriate pain control. Instead of emphasising ethical dilemmas and underlying principles, education should highlight communication skills, clarify values, and enhance the overall understanding of the healthcare system to address potential environmental conflicts.31 This approach can ultimately increase staff confidence (ie, moral courage) in constructively communicating their concerns.42 Screening tools for various emotional states, such as the DASS-21, should also be included to help individuals gain better awareness of their own psychological well-being and seek professional help if necessary. Additionally, these tools can be used to monitor emotions that might cause moral distress.

At the organisational level, efforts should be made to promote intra- and interdisciplinary communication. Poor communication, one of the five most morally distressing items, can lead to diminished quality of care, reduced job satisfaction, and poor patient outcomes.46 Ethics rounds, formal and informal discussions, and debriefing sessions regarding morally distressing cases could improve interdisciplinary communication.22 These initiatives can help promote better mutual understanding of viewpoints across disciplines and individuals.22 Participation in these events may also allow nurses to feel more empowered and experience a greater sense of decision-making autonomy.43 Finally, the establishment of formal ethical consultation services may provide support and clarification with respect to ethical dilemmas.44

At the administrative level, administrators should recognise that it is acceptable for staff to perceive moral distress; this perception is a sign of humanity and an affirmation of moral values.44 Improvements in clinical environments (eg, reduction of staff shortages, promotion of intra- and interdisciplinary collaboration, and encouragement of a safe and supported ethical climate) can help decrease moral distress.47 These measures include providing respectful feedback to staff, empowering staff to voice perceptions and emotions, and making difficult decisions in a timely manner after open discussion.48

This study revealed significant differences in moral distress among doctors, nurses, and allied health professionals in a newly established PICU in Hong Kong. Nurses had the highest moral distress scores among the three groups of PICU healthcare workers in this study and among published studies involving PICU nurses. Most areas of moral distress were related to end-of-life care and poor communication. Higher moral distress was also associated with greater depression, anxiety, and intention to quit. There is an urgent need for interventions to help healthcare workers cope with moral distress and create a more sustainable working environment.

Concept or design: WL Cheung, KL Hon, KKY Leung, WF Hui.
Acquisition of data: WL Cheung, KL Hon, KKY Leung, WF Hui.
Analysis or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

As an editor of the journal, KL Hon was not involved in the review process. Other authors have disclosed no conflicts of interest.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This research was approved by the Hong Kong Children’s Hospital Research Ethics Committee (Ref No.: HKCH-REC-2020-008) and was conducted in accordance with the Declaration of Helsinki and International Conference on Harmonisation Good Clinical Practice Guideline. All participants provided informed consent to take part in the research.

1. Jones S, Rantell K, Stevens K, et al. Outcome at 6 months after admission for pediatric intensive care: a report of a national study of pediatric intensive care units in the United Kingdom. Pediatrics 2006;118:2101-8. Crossref

2. Cremer R, Leclerc F, Lacroix J, Ploin D; GFRUP/RMEF Chronic Diseases in PICU Study Group. Children with chronic conditions in pediatric intensive care units located in predominantly French-speaking regions: prevalence and implications on rehabilitation care need and utilization. Crit Care Med 2009;37:1456-62. Crossref

3. Namachivayam P, Shann F, Shekerdemian L, et al. Three decades of pediatric intensive care: who was admitted, what happened in intensive care, and what happened afterward. Pediatr Crit Care Med 2010;11:549-55. Crossref

4. Edwards JD, Houtrow AJ, Vasilevskis EE, et al. Chronic conditions among children admitted to U.S. pediatric intensive care units: their prevalence and impact on risk for mortality and prolonged length of stay. Crit Care Med 2012;40:2196-203. Crossref

5. Ebrahim S, Singh S, Hutchison JS, et al. Adaptive behavior, functional outcomes, and quality of life outcomes of children requiring urgent ICU admission. Pediatr Crit Care Med 2013;14:10-8. Crossref

6. Bruce CR, Miller SM, Zimmerman JL. A qualitative study exploring moral distress in the ICU team: the importance of unit functionality and intrateam dynamics. Crit Care Med 2015;43:823-31. Crossref

7. Austin W. Moral distress and the contemporary plight of health professionals. HEC Forum 2012;24:27-38. Crossref

8. Jameton A. Nursing practice: the ethical issues. Englewood Cliffs (NJ): Prentice Hall Inc; 1984.

9. Garros D, Austin W, Carnevale FA. Moral distress in pediatric intensive care. JAMA Pediatr 2015;169:885-6. Crossref

10. Lusignani M, Giannì ML, Re LG, Buffon ML. Moral distress among nurses in medical, surgical and intensive-care units. J Nurs Manag 2017;25:477-85. Crossref

11. Sannino P, Giannì ML, Carini M, et al. Moral distress in the pediatric intensive care unit: an Italian study. Front Pediatr 2019;7:338. Crossref

12. Dodek PM, Wong H, Norena M, et al. Moral distress in intensive care unit professionals is associated with profession, age, and years of experience. J Crit Care 2016;31:178-82. Crossref

13. Ohayon MM, Priest RG, Guilleminault C, Caulet M. The prevalence of depressive disorders in the United Kingdom. Biol Psychiatry 1999;45:300-7.Crossref

14. de Vargas D, Dias AP. Depression prevalence in intensive care unit nursing workers: a study at hospitals in a northwestern city of São Paulo State. Rev Lat Am Enfermagem 2011;19:1114-21. Crossref

15. Poghosyan L, Clarke SP, Finlayson M, Aiken LH. Nurse burnout and quality of care: cross-national investigation in six countries. Res Nurs Health 2010;33:288-98. Crossref

16. Humphries N, Morgan K, Conry MC, McGowan Y, Montgomery A, McGee H. Quality of care and health professional burnout: narrative literature review. Int J Health Care Qual Assur 2014;27:293-307. Crossref

17. Prompahakul C, Epstein EG. Moral distress experienced by non-Western nurses: an integrative review. Nurs Ethics 2020;27:778-95. Crossref

18. Hamric AB, Borchers CT, Epstein EG. Development and testing of an instrument to measure moral distress in healthcare professionals. AJOB Prim Res 2012;3:1-9. Crossref

19. Tran TD, Tran T, Fisher J. Validation of the Depression Anxiety Stress Scales (DASS) 21 as a screening instrument for depression and anxiety in a rural community-based cohort of northern Vietnamese women. BMC Psychiatry 2013;13:24. Crossref

20. Wang K, Shi HS, Geng FL, et al. Cross-cultural validation of the Depression Anxiety Stress Scale–21 in China. Psychol Assess 2016;28:e88-100. Crossref

21. Trotochaud K, Coleman JR, Krawiecki N, McCracken C. Moral distress in pediatric healthcare providers. J Pediatr Nurs 2015;30:908-14. Crossref

22. Sauerland J, Marotta K, Peinemann MA, Berndt A, Robichaux C. Assessing and addressing moral distress and ethical climate part II: neonatal and pediatric perspectives. Dimens Crit Care Nurs 2015;34:33-46. Crossref

23. Papathanassoglou ED, Karanikola MN, Kalafati M, Giannakopoulou M, Lemonidou C, Albarran JW. Professional autonomy, collaboration with physicians, and moral distress among European intensive care nurses. Am J Crit Care 2012;21:e41-52. Crossref

24. Corley MC. Nurse moral distress: a proposed theory and research agenda. Nurs Ethics 2002;9:636-50. Crossref

25. Wilkinson JM. Moral distress in nursing practice: experience and effect. Nurs Forum 1988;23:16-29. Crossref

26. Larson CP, Dryden-Palmer KD, Gibbons C, Parshuram CS. Moral distress in PICU and neonatal ICU practitioners: a cross-sectional evaluation. Pediatr Crit Care Med 2017;18:e318-26. Crossref

27. Solomon MZ, Sellers DE, Heller KS, et al. New and lingering controversies in pediatric end-of-life care. Pediatrics 2005;116:872-83. Crossref

28. Molloy J, Evans M, Coughlin K. Moral distress in the resuscitation of extremely premature infants. Nurs Ethics 2015;22:52-63. Crossref

29. McCarthy J, Gastmans C. Moral distress: a review of the argument-based nursing ethics literature. Nurs Ethics 2015;22:131-52. Crossref

30. LaSala CA, Bjarnson D. Creating workplace environments that support moral courage. Online J Issues Nurs 2010;15:Manuscript 4. Crossref

31. Epstein EG, Delgado S. Understanding and addressing moral distress. Online J Issues Nurs 2010;15:Manuscript 1. Crossref

32. Ahmed GE, Abosamra OM. Knowledge of pediatric critical care nurses regarding evidence based guidelines for prevention of ventilator associated pneumonia (VAP). J Educ Pract 2015;6:94-101.

33. Hickey PA, Pasquali SK, Gaynor JW, et al. Critical care nursing’s impact on pediatric patient outcomes. Ann Thorac Surg 2016;102:1375-80. Crossref

34. Earle CC, Park ER, Lai B, Weeks JC, Ayanian JZ, Block S. Identifying potential indicators of the quality of end-of-life cancer care from administrative data. J Clin Oncol 2003;21:1133-8. Crossref

35. Wösten-van Asperen RM, van Gestel JP, van Grotel M, et al. PICU mortality of children with cancer admitted to pediatric intensive care unit a systematic review and meta-analysis. Crit Rev Oncol Hematol 2019;142:153-63. Crossref

36. Davies B, Clarke D, Connaughty S, et al. Caring for dying children: nurses’ experiences. Pediatr Nurs 1996;22:500-7.

37. Johnston EE, Alvarez E, Saynina O, Sanders L, Bhatia S, Chamberlain LJ. Disparities in the intensity of end-of-life care for children with cancer. Pediatrics 2017;140:e20170671. Crossref

38. Demaret P, Pettersen G, Hubert P, Teira P, Emeriaud G. The critically-ill pediatric hemato-oncology patient: epidemiology, management, and strategy of transfer to the pediatric intensive care unit. Ann Intensive Care 2012;2:14. Crossref

39. Dalton HJ, Slonim AD, Pollack MM. Multicenter outcome of pediatric oncology patients requiring intensive care. Pediatr Hematol Oncol 2003;20:643-9. Crossref

40. Zhang W, Wu X, Zhan Y, Ci L, Sun C. Moral distress and its influencing factors: a cross-sectional study in China. Nurs Ethics 2018;25:470-80. Crossref

41. Beumer CM. Innovative solutions: the effect of a workshop on reducing the experience of moral distress in an intensive care unit setting. Dimens Crit Care Nurs 2008;27:263-7. Crossref

42. Prentice T, Janvier A, Gillam L, Davis PG. Moral distress within neonatal and paediatric intensive care units: a systematic review. Arch Dis Child 2016;101:701-8. Crossref

43. Browning AM. CNE article: moral distress and psychological empowerment in critical care nurses caring for adults at end of life. Am J Crit Care 2013;22:143-51. Crossref

44. Santos RP, Garros D, Carnevale F. Difficult decisions in pediatric practice and moral distress in the intensive care unit. Rev Bras Ter Intensiva 2018;30:226-32. Crossref

45. Burston AS, Tuckett AG. Moral distress in nursing: contributing factors, outcomes and interventions. Nurs Ethics 2013;20:312-24. Crossref

46. Whitehead PB, Herbertson RK, Hamric AB, Epstein EG, Fisher JM. Moral distress among healthcare professionals: report of an institution-wide survey. J Nurs Scholarsh 2015;47:117-25. Crossref

47. Atabay G, Çangarli BG, Penbek Ş. Impact of ethical climate on moral distress revisited: multidimensional view. Nurs Ethics 2015;22:103-16. Crossref

48. Benoit DD, Jensen HI, Malmgren J, et al. Outcome in patients perceived as receiving excessive care across different ethical climates: a prospective study in 68 intensive care units in Europe and the USA. Intensive Care Med 2018;44:1039-49. Crossref