**2. Reconsidering patient safety from a social work perspective**

Patient safety is most often viewed as a technical endeavor, that is one with is focused on the identification, mitigation and response to risks and errors [22]. But beyond the technical aspect, there is also another dimension to patient safety, and one which has generally been 'hidden in plain sight'. That is that the as history shows certain groups have always been at higher risk of harm within healthcare systems and services, and those risks are not necessarily associated with the condition which is being treated, but rather with their social status.

One of the most popular ways to manage public distress over large scale failures of patient safety has been through public inquiries. Even a cursory glance at inquiries around the world show that particular groups – people with disabilities (particularly but not only those with mental illnesses), older adults, women are over-represented as the victims in these cases [23]. Such inquiries go back into the 1800s most often at that time into the treatment of patients in what where then known as 'lunatic asylums'. More recent inquiries range those into individuals who systematically murdered patients, as in the case of Harold Shipman (most of whose victims were older women) [24] to wholesale failures of governance as occurred in the case of Mid-Staffordshire Hospital in the UK, where many patients died not of medical errors but of de-hydration, that is because of a lack of human rather than technical care [25].

In a study conducted in NSW, Australia with 195 clinicians we asked who the clinicians thought were at risk of harm in the healthcare system. They identified the following groups (in no particular order): clinicians; the older adults; Indigenous peoples; immigrants – especially those with limited local language skills; people with disabilities, especially people with cognitive impairments; children and youth; patients with literacy and communication problems; people from lower socio-economic backgrounds; geographically isolated individuals; socially isolated individuals; people who are homeless; the frail and malnourished; prisoners; patients with co-morbidities and chronic illness; patients with high acuity and complex system dependence (e.g. on dialysis); those with liminal (social, physical, geographic) status; and those patients without an advocate [26].

We replicated part of this study with a survey distributed through the International Society for Quality in Healthcare (ISQUA), that is to an international audience, 15 years later in 2018. In that (still unpublished study) 413 participants from around the globe identified a very similar list of groups including: older people, including frail older people and people with cognitive impairments; children, babies, newborns, 'young people'; patients in general, 'anyone in health care', 'vulnerable patients'; people with specific, complex or co-morbid medical conditions; people with cognitive impairment, learning problems or intellectual disability; health professionals and others employed in health facilities; people from culturally or linguistically diverse backgrounds, 'ethnic minorities'; people in specific areas e.g. ICU, emergency, surgery; people with mental health problems and or substance abuse; people with physical or sensory impairment; people with limited education and or literacy; poor people, low socio-economic status, 'underserved'; people unable to communicate; people with multiple medications; indigenous people; homeless people; pregnant women; people with no advocates in health system, 'no friends or family'; people with rare conditions; and women. The risk to the most vulnerable individuals and groups, it seems, has not abated and could help explain at least in part why the overall rate of errors is not falling.

Why are these groups at higher risk of harm? There are several inter-related issues that have been identified. As we noted in a series of literature reviews we conducted on this very issue Travaglia, Debono ([27], p. 6) *"Cascade iatrogenesis [28, 29], where one error leads to others, may be an additional risk for these groups, as may the effects of diagnostic overshadowing [30, 31], where a condition (e.g. a brain tumor) remains undiagnosed because the clinician attributes the manifestations to another (e.g. mental) illness [32]. Understanding the complex and inter-related social factors that increase ill-health in individuals, groups and communities provides a starting point for understanding why, if, when and how people access and utilize healthcare, how ill they are when they do so, and how these factors might affect their susceptibility to medical errors and adverse events [33]."*

One way of diving deeper into understanding why these particular groups are at heightened risk of harm is through the lens of social epidemiology. Social epidemiology is *"… distinguished by its insistence of explicitly investigating social determinants of population distributions of health, disease and wellbeing, rather than treating such determinants as mere background to biomedical phenomena"* ([34], p. 693).

Kreiger's (2001) framework takes into account the: biological expressions of social inequality (that is the embodiment of experiences of socio-economic inequality, which often result in poorer health across the lifespan for these groups); impact of discrimination; eco-social theories of disease distribution (i.e. population level patterns of health, disease and wellbeing); gender, sexism and sex; human rights and social justice; life-course perspectives; poverty, deprivation and social exclusion; psychosocial epidemiology (the health damaging effects of psychosocial stress); 'race'/ ethnicity and racism; sexualities and heterosexism; social and cultural perspectives

of health; social class and socioeconomic status; social determinants of health; effects of social inequality and inequity in health; social production of disease/political economy of health; social production of scientific knowledge; stress; and theories of disease distribution (across time and space) [34]. We have added ableism to the list Travaglia et al. [27].

In 2019 we undertook literature reviews looking at the risk of harm for eight vulnerable groups: people from culturally and linguistically diverse (CALD) backgrounds; older people; lesbian, gay, bisexual, transgender, queer and inter-sex (LGBTQI+) people; First Nations peoples; people who are homeless; people living in rural and remote communities; prisoners; and people with intellectual disabilities.

*"Some studies found significantly higher risk of harm among the vulnerable groups. Instances of suboptimal care included: misdiagnosis; hospital-acquired infections; less active or inappropriate treatment; preventable readmissions; less frequent health screening; inappropriate prescribing; and poor communication between patients and providers with potentially damaging consequences. In addition to their underlying health needs, people in remote regions may experience less effective healthcare due to under-resourced services, distance to specialist care, or high turnover of health professionals. Some studies reported inadequate recognition and treatment of comorbid conditions, e.g. among people with intellectual disabilities.*

*Our researched showed the intersectional nature of the risk to patient safety, which multiplies with the number of vulnerable groups of which the individual is a member.*

*Studies highlighted the intersectional nature of patient safety, where individuals experience poor treatment because they belong to two or more vulnerable groups, e.g. First Nations people living in remote locations, or people with intellectual disability in prison. This compounds their vulnerability, increasing the risk within health systems. Studies also found some disadvantaged people reluctant to access health services, exacerbating their already poor health, due to negative previous experiences or fear of discrimination or disrespect" ([27], pp. 3-4).*

While the understanding of the causes of ill health from a social perspective goes back decades and is very well established, this approach has not widely been employed in understanding the causes of iatrogenic harm/illness. However looking at harm from this perspective provides additional insights not only into the causes but also into potential strategies to address these harm. This is where social work leadership could come to the fore.
