Parity of Esteem and Systems Thinking. A Theory Informed Qualitative Synthesis


 Background: UK policy recognises the interdependence of physical and mental health, but the failure to define Parity of Esteem (PoE) in any depth, challenges its implementation. There are at least two shortcomings in current PoE implementation; the lack of specificity in its operational components and progress indicators and the sub-optimal ‘siloed’ approach to its redress. Methods: This is a descriptive qualitative study employing 27 qualitative interviews and the World Health Organisation (WHO) system domains as a framework for the thematic analysis. Results: Examining the current strengths and challenges of systems in mental and physical healthcare using the WHO domains and macro, meso and micro levels, identifies specific actions to redress inequity between mental and physical health provision. Conclusion: The evidence suggests that moving PoE from rhetoric towards reality requires new configurations with a systems orientation, which uses macro, meso and micro levels to analyse and understand the complexity of relations within and between domain levels and reorienting funding, training and measurement. This requires embedding new competencies, infrastructures and practices within an effective learning healthcare system.


Introduction
There are large inequalities between the physical health of people with and without mental health conditions. Inequalities manifest as a signi cantly higher risk of dying earlier from preventable physical illnesses among those diagnosed with mental health di culties and pervasive barriers to accessing healthcare 1,2,3,4,5,6,7 . Achieving parity between physical and mental health is a national priority, but there are many challenges to overcome and the area is fraught with bureaucratic, regulatory and structural complexity 8, 9 .
The concept of parity of esteem (PoE) in healthcare began with its enshrinement in USA law under the Mental Health and Parity Act (MHPA, 1996), with further application of the concept occurring through the Mental Health Parity and Addiction Equity Act (MHPAEA, 2008) 10 . The UK coalition government introduced PoE in 2011, in the White Paper 'No Health without Mental Health', promoting equality between physical and mental health care 11 (p. 64, Para. 7.1). This paper recognised the interdependence of physical and mental health, but failed to de ne PoE in any depth, challenging its implementation. In 2012, the Health and Social Care Act made parity a requirement in law, resulting in an NHS Mandate in 2015/16 12 (p.16) and ensuing policy documents and professional guidance 13,14,15 . There remains however, opacity regarding its de nition and its use in health and social care settings.
One way of reframing the numerous conversations and complexity surrounding parity of esteem is to employ systems thinking. The utility of a systems-level approach may overcome at least two shortcomings in current PoE implementation: i) the lack of speci city in its operational components and progress indicators and ii), the sub-optimal 'siloed' approach to its redress.
The World Health Organisation [WHO] framework outlines the core building blocks necessary for developing and sustaining effective health systems 16 . The six critical domains are; nancing; service delivery; the health workforce; information; medicines and technology; and leadership and governance. In theory, improvements to the quality, responsiveness and coverage of all these elements should lead to improved health equity and parity. This approach enables a shift in thinking from a more linear and segmented view towards more complex systems founded on synergistic components 17,18,19 .
This paper frames PoE within this systems-based approach to examine the current strengths and challenges of this concept in mental healthcare and identify speci c actions to redress inequity between mental and physical health provision.

Methods
This study conforms to COREQ guidelines, using a descriptive qualitative design, informed by the WHO health systems framework 16 .

Sample and recruitment
Study participants were local, regional or national key informants in mental health policy and practice (n=27). Purposive and snowballing sampling identi ed local and national key informants. The basis of selection rested on the premise that participants possessed a critical understanding of parity of esteem within and across sectors.
Recruitment occurred between June 2020 and December 2020, initially through an advert on the local healthcare intranet site and via community groups within Greater Manchester. Recruitment snowballed through social media and participants' own contacts to form a national lens, ending when reaching data saturation.
Twenty-seven participants self-selected by agreeing to an interview and providing their contact details.
Participants included mental health care providers, physical health care providers, policy-makers, service commissioners, mental health charity workers and political party members (see Table 1). conformed to the ethical norms and standards in the Declaration of Helsinki. Participants were given the information sheet and contacted after two weeks, leaving them time to consider the study and ask questions. The consent procedure discussed the study, asking participants if they had any further questions before obtaining informed consent to participate. Participants consented to the use and analysis of data for publication and teaching purposes. Anonymising participant transcripts (PT) assured anonymity. All data was password protected and kept on encrypted servers. Deletion of interview recordings occurred immediately post-transcription. The datasets generated and/or analysed during the current study are not publicly available because they contain identi able data.

Results
Analysis used the domains from the WHO systems framework 16 as a lens through which to view data and develop themes. It also applied macro, meso and micro levels to identify barriers and facilitators. Table 2 illustrates the various themes mapping onto system domains and levels. Domains: Barriers and enablers of PoE at macro, meso and micro levels

Leadership and Governance
Key themes for this domain included a) policy and action plans on parity and b) shared decision-making. Policy and action plans at the macro level for parity of esteem appear to have created barriers due to a lack of clarity and de nition in the ways that parity may be achieved 20,21 . Broad enablers for PoE at the macro level included "campaigning for mental health funding" (PT.19, p.3), and "making sure mental health problems have equivalent policy priority"(PT. 18, p.2), but the method of enactment of these initiatives required meso and micro level thinking. Speci cally, Participants argued that setting equality standards at the macro level such as everyone having a maximum of a two-week wait prevented shared decision-making, creating barriers because it worked against people experiencing a mental health crisis, tending to treat all mental health conditions as the same with little consideration for diversity, or complexity. "Why is it that people with psychosis, or a psychotic crisis, should expect the equivalent, or parity, if you like, of access, as someone with […] cancer?"(PT.1, p.12).
Shared decision-making inevitably links to patient-centred care, and equity in this context meant having the autonomy to build a relationship and respond to a patient as a person situated in a particular psychosocial and economic environment. The majority of this activity occurred at the meso and micro level. "To understand that what generates good diabetic care, is often understanding psychological and social contexts in which people have their di culties" (PT.1, p. 9).
Financing Key themes for this domain included a) funding for services, b) targets and incentives and c) the mental health estate.
All participants mentioned the central role of funding; suggesting that a system was constrained by inequitable funding was both a challenge to and a manifestation of a lack of PoE. Participants perceived that moving funding around to prioritise one area over another perpetuated mental and physical health inequalities, inadvertently setting the two elds up as nancial competitors rather than an integrated healthcare goal. "In a resource constrained system, if we want to give more to someone else we have to take it from one area and move it to another"(PT. 27, p. 14).
Integrated care systems occur at meso levels but remain heavly dependent on funding from the macro level, with mental health appearing secondary to physical health "there's always going to be people with what appear to be hugely signi cant health problems physically that "trump" mental health problems" (PT. 27, p.5). Participants mentioned further local tensions in resource allocation and funding for acute and mental health services, emanating from this inequity. "[…] each ICS (integrated care system) has a sort of capital spending limit, which has to be shared between its acute hospitals and its mental health trusts. Mental health is always a relatively small player" (PT. 6, p.4). Participants believed that this Cartesian vision of mental and physical health limited the ability to gain larger amounts of funding from the macro level.
Participants also discussed the inadequacy of mental health funding for children and adolescents, suggesting local and whole system inequalities maybe further exacerbated in some populations. "What we focus on(young people and mental health)is so sort of underfunded and under-prioritised compared to the rest of the healthcare system" (PT.12, p.3).
Participants perceived a need for adequate resourcing at meso levels to meet policy and service targets "the need is to focus resources and provision to meet those targets. I think the challenge with it is that you have to have adequate resources to do that" (PT. 21, p. 10).
For some participants targets added a new level of complexity that risked whole system improvement. "The danger with that is we get narrowly focused on how many people are accessing psychological therapies, and a more encompassing measure looks at people who maybe don't want to access psychological therapies, but do still want support" (PT.2, p. 12). For many participants in the current study, targets did not appear to address provision for people with complex mental and physical health problems, resulting in a lack of parity. "There are waiting time targets for psychological therapies, but it seems that not really that much is done around those people who fall through the gaps and who are too complex" (PT.2, p, 7). The complexity of conditions appeared to be an area receiving very little consideration at macro level when allocating funding."The other big test of the complexity of parity of esteem, in a sense, is that people don't have one problem. So people don't just have a psychological di culty, or they don't just have diabetes, they often have multiple problems"(PT. 1, p. 14). At worst, participants suggested that targets might lead to a reduction of total effort at the meso level "sometimes what happens then is people only focus on the targets, rather than the bigger picture"(PT. 5, p.8).
A considerable challenge involves clearly identifying indicators for benchmarking PoE progress "the indicators are going to act as this kind of galvanising force around an area that the targets are going to do that so you have to be very speci c about what it is you want change" (PT. 8, pp.6-7).
Participants perceived that a lack of funding at the macro level affected both the physical environment of mental health services at meso level and the availability and quality of care for people with mental health problems at micro level. A lack of investment in physical infrastructure affected mental health care creating "facilities that are not t for purpose, with people still in dormitories and wards that are not acceptable in this day and age"(PT. 6, p. 7). A concomitant impact on patient wellbeing and long-term recovery was evident."Often in emergency departments there will be a particular area or a particular room where someone who is experiencing a mental health crisis, or presenting with a primary mental health need, will be placed. Often those places aren't very well looked after; they're not very welcoming, they're not very therapeutic."(PT. 25, p. 5).

Service Delivery
Key themes in the service delivery domain were a) Access to quality care, b) Integrated care and collaboration, c) Informal caregiver involvement and d) Person-centred care.
Macro level enablers for participants included "making sure mental health problems have equivalent clinical priority"(PT. 18, p.2).
Crisis services in particular exhibited many barriers to support and care at a meso level. "How easy it is for people with mental health problems and their families to get support when somebody's in crisis. How responsive is that service? The answer is it isn't" (PT 11, p.5).
Participants discussed the accessibility of formal services and the sub-optimal role that GPs may play as gatekeepers to this system "It's no criticism of GPs, but Access Points rely heavily on people going through GPs to access the system." (PT 3, p.5). Participants offered potential solutions such as meso level enablers involving multiple access points and better health care integration and support within communities, potentially reducing inequalities in referrals for people with mental health problems and relieving the burden on GPs by employing more community-based health care for mild to moderate conditions. "There needs to be equity of access once you pass the threshold and therein lies the problem, because access for someone with a mental health problem, even to a service which has parity of resource, has to be different to someone with a physical health problem."(PT 27, p. 6). "[…] the positive change that needs to take place, that would improve mental health would be, better integration with the community. That actually that applies to all long-term conditions."(PT 8, pp. 5-6). Participants discussed involving families and formal carers as facilitators for developing greater levels of acceptability. "You've got to start with the service users, their families and the professionals that are directly involved in the coordination and delivery of their care and support."(PT 13, p. 14). Without the support of families and formal caregivers, professionals may struggle to deliver acceptable, effective and equitable care, but this itself raises challenges at the micro level in terms of language used, clinical culture and power differentials.
Barriers to patient-centred care involved time and pressure on services "time is an important factor because of increasing service demands and pressure on clinical services. Sadly, even with the best will, we don't have the time to sometimes delve into a little more detail." (PT. 24, p.5). Without the time to explore with patients and their carers important details and inclusion in the patient professional encounter then 'no decision about me without me' 22 (p.3) is unlikely to occur. Some participants discussed person-centred care as both a facilitator and barrier to parity. "Making sure that the patient is able to tell you more about their needs than anything else"(PT16, p.6). However, "With mental health I think everyone is a bit more fearful; they remove themselves, they maybe don't want to enter in a discussion" (PT. 23, p.6). It would appear that for parity to occur in choice and decision-making there needs to be more training in facilitating individuals, families and carers to become part of the decisionmaking environment.

Workforce
Key themes for this domain were a) Education, training and continuous professional development (CPD), and b) Sta ng numbers.
Participants perceived the impact of lack of funding on sta ng numbers at macro level as a challenge to service delivery at the meso level"When patients come into ED, they're often waiting for the mental health liaison team. They're quite often short-staffed, underfunded" (PT. 23, p. 3). They felt that a lack of resources had a direct impact on staff themselves, particularly in early intervention and crisis care services, which then exerted an effect at the micro level: "the staff are under so much pressure, they have no resources and not enough staff" (PT. 22, p. 11).
"When someone's in a mental health crisis, they should not be left to not know where to go. Or to not have support because there aren't enough resources or the service isn't open" (PT. 2, p. 7).

Sta ng was an issue throughout participant interviews, particularly at the meso level for
The workforce is a substantial resource within the services, and arguably, it is important to keep abreast of new evidence and ensure holistic care is prominent. Some participants highlighted a siloing of education and training which focused mostly on physical health "If I'm trained as a physician I'm trained in physical health. I learn to be blind to mental health problems"(PT. 27, p.4). Cartesian duality appeared rei ed by education "you know we're de nitely trained in splitting people's bodies and minds."(PT1. P, 9).
This disjointed training at the meso level disabled healthcare practitioners, affecting service delivery at the micro level, because there are barriers to effective communication and asking people about their mental health. "Mental health nurses, don't feel as con dent with physical healthcare as we could, because we haven't been trained to the level that we need. The reverse is true for acute care nurses, they don't get any mental health training and they're scared" (PT. 10, p. 4).

Information and Research
Key themes for this domain were; a) Access to reliable data; b) Measurement and benchmarking; and c) Parity in research.
Acknowledgement came from participants regarding the lack of parity between physical and mental health research "you'd want to see some parity in mental health research. Of course, it's way behind other health areas." (PT. 14, pp. 7-8). Reducing barriers at the macro level would facilitate more equitable meso level advances. Whilst broad macro level enablers for participants included "making sure mental health problems have equivalent research and funding priority"(PT. 18, p.2), more speci city was evident in some participants responses.
Participants discussed meso level barriers to accessing reliable data because of variations in coding and collecting data. "Quality of the data is still really poor and the way that different NHS trusts collect the data, there's just a huge amount of variation. How can you assess the effectiveness of the system, the quality of care, if rst of all you're not collecting the right data?"(PT. 12, p. 14). Enablers involved standardising data aiming towards quality improvement and "really understanding the data and what you're measuring, really understanding the change cycle that you're implementing, and then sustainability about the metrics that you're collecting, to measure it." (PT. 21, p.15). Another participant suggested that there was a pressing need "to invest more time and energy and effort and research into what patient reported outcomes (PROMs) would be of value to people." (PT. 24, p. 6).
Being able to judge the e cacy of a system and need for resources becomes challenging if the data is unreliable. It also makes requests for further resources di cult.
Within healthcare, benchmarking links to targets, incentives and metrics and usually classi ed into one of four categories: productivity, quality, time and cost-related. This allegedly establishes standards of excellence and improves services and quality 23 . Participants perceived that benchmarking was important "how are we doing on this indicator, compared to the national average? I would do it more by benchmarking the key indicators."(PT. 14, p. 8). This also underlined considerable operational complexity: "How you measure what the outcomes are for something like mortality? Do you then break it down for different mental health conditions?" (PT. 3, p.11).
At the meso level, participants gave positive examples of enablers such as Quality Outcomes Frameworks (QOFs) involving integrated care addressing physical and mental health. "They access a dietician, get to the optician, see the podiatrist, and go for their regular physical health checks. They should have an ECG and bloods taken at regular intervals at least once a year. People with bipolar disorder and schizophrenia should have that in general practice. There was a QoF around that, but that's been removed" (PT.17, p.7). While participants perceived incentivising physical and mental health through the introduction of yearly health checks at the meso level as an enabler, the potential for the miscoordination or instability of these approaches was a prominent barrier.

Technologies and Medical Products
Key themes for this domain were a) Shared information systems and b) Treatment choice.
This domain was least mentioned. Participants who did discuss PoE under this domain typically perceived barriers to care emanating from a lack of shared information systems at the meso level. This occurred primarily when they were attempting to support inpatients with mental health care. In exible interpretations of con dentiality and data protection appeared to present speci c barriers. For example, blocking access of healthcare staff to scans and appointments sent to the patient's home address, even though they were in-patients in a mental health unit in the same hospital grounds: "They have a system within the main hospital where all the CT scans are reported and actually held electronically, but we don't have access. We have tried to get access and we have failed. The radiologists are telling us, 'we cannot give you access to the system, we cannot email you the result, and you have to physically come'. So, we have to leave our building, walk across the hospital site, go to another building and get the scan result. When my junior doctor arrived there they said, 'oh no, we are not giving it to you'.
So, they can't email it to us, they can't physically give it to us, they can't put it on the system, so I'm not really sure how we're supposed to access it." (PT17, p.4).
Other participants discussed barriers to treatment choices resulting from perceptions around mental health at the micro level "I would always discus treatments and the choice and the availability of different treatments to a patient and, you know, involve, them […]. In mental health, for some reason, a different idea is that they can't make that decision or they shouldn't have that decision" (PT. 23, p.9). This participant appeared to suggest that professional or institutional discrimination might result in patients with mental health problems ultimately having less choice and by implication iniquitous or sub-optimal care standards. "They've not had the same choice in treatment, or […] I'm not going to Steri-Strip…I'm not going to suture them because, you know, they've self-harmed, they'll probably rip it out and self-harm again, so it's not really worth it." (PT. 22, p.3-4).

Discussion
The aim of this paper was to frame PoE within a systems approach in order to deconstruct it into clearly de ned domains and system levels and describe facilitators and challenges. Delineations between macro, meso and micro levels are not always clear, for example, in the domain of service delivery, training de ciencies are meso level but could also be micro level because this affects patients. Another argument is that it could be macro level because policy decisions can alter healthcare curricula or CPD requirements. Table 2 displaying the domains and levels may not be comprehensive, but for this study, there appear to be more barriers at macro and meso and micro levels across domains than enablers for PoE.
The domain for Leadership and governance exhibited enablers at macro and micro level through policy and guidance, but simultaneously participants viewed these actions as barriers because of a lack of clarity. There was no evidence of leadership and governance at meso level, leaving decisions about PoE to individuals. Setting standards at the macro level such as a two-week wait for all conditions created inequity because of the failure to acknowledge difference and complexity between and within conditions and people. Shared decision-making at the micro level acted as an enabler for parity, but a barrier at the micro level, because of a lack of time, creating a negative impact in service delivery.
Importantly, nancing exerted an impact across all domains. Participants suggest that a lack of funding and resources creates inconsistencies in resource allocation, care quality and outcome measurement.
The failure to consider complexity remains unaddressed, affecting multiple health system domains and implementation levels. Governments rightly aim to achieve real value-for-money through public investment in service delivery. However, the lack of recognition concerning the complexity of the environment and the impacts of policy changes on service delivery appears to exclude equitable care improvements and measurement. For mental health, outcome measures need to be valid, reliable, sensitive to change, comparable across a range of service users and meaningful for both clinicians, patients and their supporters 24,25 . The main problem that the participants in this study identify is that implementing outcomes reactively maintains contractual and funding requirements rather than enacting well-de ned and appropriate goals, which may more readily lead to equity.
At micro-level, enablers may include the introduction of more meaningful outcome measures re ecting improvement in social and occupational functioning, mental health symptoms and distress, and physical health and wellbeing. This intention would strengthen progress in the information and research system domains because whilst outcomes measurement offers valuable information on the impact and effectiveness of a service, it also presents its own challenges. For example, valid outcome measurement may be dependent on the availability of reliable evidence and data capture tools and affected by the complexity of systems at meso level.
The Service delivery domain focused heavily on 'Access to care', which demonstrates intrinsic and direct links to outcomes measurement and well-de ned goals instigated at the macro and meso levels 26 .
Access to quality care is currently di cult to measure with certainty because it is a theoretical concept encompassing different aspects and dependent on the exact de nition and the context of measurement 27 . Enablers at the meso level may include multiple access points to support services in the community, extending the reach of community support services and facilitating access to crisis care before symptoms or behaviour becomes unmanageable. These initiatives may present as quanti able improvements at the micro level through continual review, follow-up with patients and family and caregiver support to prevent escalation of symptoms. Research suggests that including family members in decision-making and choices is a positive approach to mental health, as long as agreed by all 22,28,29 .
Within the Workforce domain, the siloing of education and the creation a Cartesian divide appears in urgent need of address, leading from a macro-level reorientation of training, education and CPD focused on the holistic body. This has potential to initiate gains in multiple system domains, including the Technology and Medical Products domain at the micro level where stereotypical perceptions about mental health may still be limiting treatment choices. It also affects patient-centred care at the micro level in the Service delivery domain. The nancing domain, in itself will directly affect the workforce domain, dictating sta ng levels at the meso level and resources, and the development or otherwise of a supportive workforce culture.
The concept of a learning health system (LHS), in which an organisation builds knowledge or evidence, embeds quality improvement practices as standard practice, actively engages patients and family members and supports further learning 30 , offers a potentially fruitful approach to building parity across multiple interacting domains and levels of healthcare. Taking these four elements of the learning healthcare system and using them in a PoE focused systems-based approach may offer important traction.

Study limitations
The sample limited collection of data during lockdown in 2020/2021. Our assessment of macro, meso, micro levels within the study represents a professional perspective, and future studies may garner a more comprehensive insight from everyone involved across different settings and levels.

Conclusion
Moving PoE from rhetoric to reality suggests the need for new con gurations that have a systems orientation. This includes new approaches from macro to micro levels, which focus on reorienting funding, training and measurement and that are oriented towards understanding the complexity of relationships within and between domains and levels. This requires ongoing development, monitoring, evaluation and may most usefully occur by adopting a whole systems approach and by embedding this principle within an effective learning healthcare system.

Declarations
Ethical approval and consent to participate Availability of data and materials The qualitative data generated and analysed during the current study are not publicly available because they contain personal data, for example, job roles, place of work, colleague names, family names and circumstances, which could identify the participant. Redacted data are available from the corresponding author on reasonable request.