Applying a ‘whole systems’ approach to infection prevention

Applying a ‘whole systems’ approach to infection prevention & control in primary health care clinics in South Africa Critical reflections from the Umoya omuhle project 1000 UK 1100 SA Wednesday 28 April 2021 Funded by the ESRC/AMR Cross-Council Initiative and The Bloomsbury SET

Welcome and introduction Alison Grant Karina Kielmann London School of Hygiene & Tropical Medicine and Africa Health Research Institute Queen Margaret University

Applying a ‘whole systems’ approach to infection prevention & control in primary health care clinics in South Africa

RQ3 IPC: Clinic design, work practices RQ4 RQ2 HCW and patient ideas and practices of IPC TB transmission in facilities RQ5 Systems interventions to improve IPC RQ1: Analysis of IPC policy RQ6 Mathematical modelling: effect on transmission? RQ7 Economic impact of reduced transmission

Why is IPC of interest to systems and social science researchers? Growing emphasis on HCAI - clinics as ‘safe’ yet permeable ‘risky’ spaces Compromised IPC generally seen as a 'behavioural problem' of non-compliance rather than a systems problem; Complex - multiple components of IPC from buildings, space, technology through to less visible drivers of workflow, air flow, and flow of people is organised within given space

Relevance for other infectious diseases Emergence of other ID outbreaks including COVID-19 highlight need for better IPC Creating and sustaining a ‘culture of safety’ in resourcepoor health systems experiencing ID outbreaks Health care settings at frontline of ‘containment’ and response strategies; HCW balance global standards of IPC with locally specific constraints Challenges of implementing systems change within context of discursive fragmentation in policy context Dramowski, Woods, and Mehtar (2020)

Structure of Webinar 1 2 3 4 5 6 7 Understanding complexity through clinic Anna Voce ethnography of TB IPC Assessing material and social infrastructure of Hayley TB IPC MacGregor Helen Schneider Discussant comments Thilo Govender Open discussion Generating evidence through system Karin Diaconu dynamics modelling Justin Parkhurst Designing and costing whole systems Fiammetta interventions Bozzani Alastair Ager Discussant comments Gerald Bloom 8 9 Open discussion Wrap-up Janet Seeley Chris Colvin Anna Vassall Karina Kielmann, Chris Colvin, & Alison Grant

Housekeeping This session will be recorded and made available on the Umoya omuhle websites at LSHTM and QMU. Please mute during presentations Please also feel free to use the chat function to put your questions in writing During open discussion: please unmute, briefly introduce yourself, and ask questions or provide comment

Unravelling complexity Clinic ethnography of TB IPC Implementation in South African Public Sector Primary Care Health Facilities Anna Voce School of Nursing and Public Health University of KwaZulu-Natal

Background Perspective adopted by Umoya omuhle - TB IPC studied as a ‘complex’ intervention, dependent on: A number of interacting, multi-synergistic components, across the levels of the health system; Changes to the organisation of care, use of space, and to administrative processes; Changes in perception and behaviour, from different categories of health care workers, and from patients; Different types of professional activity from within and across the levels of the health system Flexibility and adaptations to local care contexts

Aim Analysis of TB IPC implementation within the micro-system of primary care health facilities Understand TB IPC practices in context Service delivery environment Administrative systems and workflows Organisation of care Patient flows Organisational structures, processes and culture

Methods Case study approach, adopting immersive ethnographic methods Data generated through: Structured and unstructured observations ( 80 hours per clinic) Formal interviews and informal conversations - with healthcare managers, healthcare workers, clinic support staff, and patients Patient flow mapping 3-step thematic data analysis Thick case descriptions Within case analysis Across case analysis

Results: The main waiting area – a lens on the micro-system Conditions in the main waiting area embedded in the Integrated Clinical Services Model of care Appointment scheduling (inefficiently implemented) Pre-appointment retrieval of clinical records (not done) Single administrative point (incomplete interpretation) Re-organisation of patient flow based on streams of care (inefficiently implemented) Analysis of system challenges Not actual clinic – illustrative only Compromised TB IPC in main waiting area: Long waiting times in over-crowded waiting area, with possible extended exposure to people with active TB – exacerbated by poor ventilation, no cough triaging, no respiratory separation, no evidence of respiratory hygiene, poor use of PPE - Technical expertise for policy translation to local context Organisational structure: hierarchical Organisational culture: compliance Differentiation of management roles within and between system levels - Team synergies - Adaptive learning capacity and innovation

Implications for Health Systems Researchers Multiple complexities Multiple ways of knowing Build ability to make paradigmatic choices

Assessing material and social infrastructure Hayley MacGregor Karina Kielmann Institute of Development Studies Queen Margaret University

Infrastructure Generally refers to structural and logistical components of health facilities, e.g. “the total of physical, technical and organizational components or assets that are prerequisites for the delivery of health care services”(1) Infrastructures are not just material entities—they are also practiced and relational (2); “embedded” within established networks, relationships, dynamics, cultures or ‘social’ infrastructures 1. Scholz, S. et al (2015) Rapid assessment of infrastructure of primary health care facilities – a relevant instrument for health care systems management. BMC Health Serv Res 15, 183 2. Starr, SL (1999) The Ethnography of Infrastructure. American Behavioural Scientist

(WHO 2016) IPC: A set of ‘control levers’ to reduce risk for staff and patients by reducing exposure and transmission in clinics; made visible in standard practices e.g. masks, gloves, guidelines, windows, air extraction WHO supports ‘multi-modal strategies’ for implementation of IPC as “the way to achieve the system change, climate and behaviour that supports IPC progress and, ultimately, the measurable impact” for patients and HCW Limited understanding of the ‘enabling environment’: focus on material components while social infrastructure is less visible, hence taken-for-granted

Ventilation, Flow and PPE Interdisciplinary approach Quantitative measurements of air exchange in different sizes of rooms with different degrees of ventilation; measurement of congregation of people in different sizes of rooms, at different times of the day; calculation of the risk of transmission associated with different scenarios. Qualitative research: Interviews and observations of how staff and patients use space; exploration of reasons for lack of PPE use and ventilation of spaces; examination of the organisation of care and clinics logistics and the effect on patient flow; attention to where bottlenecks for congregation occur and why. Clinic design: interviews with architects and engineers about design priorities and processes; exploration of the challenges associated with retrofits changes

A whole systems approach Ventilation, congregation and PPE approached as separate elements in IPC assessments, with an emphasis on the material infrastructures A whole systems analysis shows the interconnections in practice between factors determining the extent of ventilation and congregation The social and material infrastructures associated with clinics are closely interconnected The interactions of both of these elements affect risk of transmission and what is possible to achieve in terms of an enabling/disabling environment for IPC

Re-thinking the ‘enabling environment’ for IPC Metrics and targets (and measurement thereof) make certain practices more visible at expense of implicit, intangible infrastructure A wider understanding of social infrastructure against the backdrop of inherited infrastructural inequalities is necessary Synergy/tensions with policies such as Ideal Clinic can enable/disable wider change e.g. in appointment systems, queue management, organisation of care into patients streams or differentiation of care Material and social infrastructure are mutually constituted

Helen Schneider Thilo Govender Janet Seeley University of the Western Cape, Cape Town, South Africa King Dininzulu Hospital, Durban, South Africa London School of Hygiene & Tropical Medicine and Africa Health Research Institute Discussants Facilitator

Discussion Unravelling complexity in a complex system

Using System Dynamics Modelling in Umoya omuhle Potential gains and losses for whole system framings Karin Diaconu Justin Parkhurst Queen Margaret University The London School of Economics and Political Science

Application in HPSR Recognition that current modelling approaches are not capturing sufficient complexity: Focus principally on one intervention Depict minimal interaction among health system components If we agree that more complexity needs to be captured, we agree to move towards “health system modelling” SDM is one method that can help with that Self-organization Non-linearity Complexity science approaches Networks Complex adaptive systems “System Dynamics is a computer-aided approach for strategy and policy design. It uses simulation modeling based on feedback systems theory and is an analytical approach that complements systems thinking. It applies to dynamic problems arising in complex social, managerial, economic, or ecological systems — literally any dynamic systems characterized by interdependence, mutual interaction, information feedback, and circular causality.” (SDM Society) Invented by Jay Forrester in the 1950s but further refined by others: Sterman, Vennix, Rouwette, Hovmand, Homer

Application within the project Two workshops (Durban, 2019) with staff from National and Provincial levels (Western Cape and KwaZulu-Natal) Policy workshop (n 9) : policymakers with expertise in TB, primary care, financing, information systems, medicine availability Practitioner workshop (n 15): patient advocates, health care staff and managers, program leadership, architecture and infrastructure specialists

Intervention targets and mechanism considerations Intervention targets and core mechanisms Number of people using the clinic Medication distribution outside clinics for chronic illness, queue management systems, appointment systems Ventilation or safety of clinic spaces 3 interventions focused on window/door opening, retrofitting spaces, installation of UVGI Personal protection measures Surgical mask wearing for patients & N95 respirators wearing for staff Other parts of the mechanism Consultations: Ensuring buy-in of patients and stakeholders at different system levels Alignment and coordination: Integrating mechanisms within existing systems as far as possible, e.g., for M&E and training and supervision Feasibility consideration: Including elements which address clinic or system level constraints to implementation and scale-up

Reflections and application in Umoya omuhle SDM is not purely objective or mechanical - tension with other methods and approaches Modelling team and participants negotiate where to focus for intervention targets and mechanisms and this is simulated: Assumptions bring tensions about what level of detail is relevant or not Modelling health system Critically interrogate and simulate system behavior Modelers distilled information from primary RQ1-4 data and GMB, but bias may be introduced: SD modelers have to move from complex and very detailed models to key dynamics in line with goals of exercise Identify system dynamics and intervention strategies Capture diverse perspectives and evidence around boundary object Unpack and understand interactions Tension over what is modelled Boundary: IPC only or broader system? Participants: whose views are included and what are their goals and views? Time-horizon: both retrospective and prospective Group model building workshops with policy makers, practitioners, advocates and researchers; GMB forces bounding of problem Evidence input by presentations from Umoya to participant group – and embedded researchers into groups to probe and challenge, but model is negotiated

Evidence use for policy in relation to SDM SDM as a ‘knowledge translation’ device As a tool to manage complexity and provide insight into policy actions Can provide insights from existing research felt to be policy relevant – exploring how it serves complex needs; Can follow a more exploratory process by which evidence is sought and considered in terms of how it serves policy needs and goals. Constructivist perspective – SDM as a tool that can be formalized within planning processes. Planning processes (and tools utilised) can shape or define: The problematisations that are seen as relevant; The solution sets held to be options; The meaning and values of elements of the policy; The representation of particular stakeholders and concerns.

Policy scholars’ perspectives on evidence use Bounded rationality – applies to decision makers and to producers of evidence Decision makers – limited time, incomplete information, unclear competing agendas; Evidence providers – Fleck (1935)’s concepts of ‘thought collectives’ - impose artificial simplification of scientific knowledge in attempts to summarise and provide evidence to others. (see Botterill and Hindmoor 2012 Policy Studies) Construction of problem and solution sets Evidence utilisation can formalise how problems are understood, which solutions are seen (and seen to be relevant), and which goals are pursued Representation and accountability Those who participate in the processes of evidence use will have particular priorities and understandings – can reflect on representation and accountability within processes.

Using a participatory approach to design and cost complex interventions Fiammetta Bozzani London School of Hygiene & Tropical Medicine

A novel approach to costing Costing framework developed alongside SDM, recognising that: interventions in our package are already variously implemented in some form because benefits are established (e.g. opening windows, wearing masks) current implementation is suboptimal and costing ‘business as usual’ activities is not going to achieve intended effectiveness targets relevant behaviours have to be contextualised within broader organisational culture, processes and system constraints, but no standard method to synthesise this information alongside economic analyses What is new? Enablers (contextual factors) identified with experts and practitioners during SDM workshops Package of interventions and enablers included in economic analysis reflects full opportunity cost of achieving control of TB transmission in clinics

Economic evaluation methods 1. SDM workshop 1 Identify interventions (impact feasibility) and specific activities to cost Identify enablers (improve feasibility) 2. Micro-costing – identify input prices and quantities Core intervention activities, enablers, consequential TB and HIV care Sources: Published literature, current research, quotes from suppliers 3. Cost model parameters validation Monthly ‘drop-in’ calls with SDM workshop participants SDM workshop 2 – present preliminary results, make changes for final estimates 4. Attach unit costs to TB transmission model outputs to generate cost-effectiveness ratios

Process outcomes and limitations Enablers can represent 90% of total costs (e.g. introducing appointments) Regular contact with GMB participants helps validate assumptions although participants may have ‘self-selected’ on drop-in validation calls Enablers tend to be above service level high-level activities, difficult to allocate to specific interventions activities that fall outside the remit of the health sector SDM-informed costing can highlight trade-offs between interventions that are more or less reliant on enablers (and at what level these operate) It provides information on intervention feasibility quantitatively within the ICER It provides qualitative evidence on enablers that are difficult to measure It is not best suited inform the choice of a cost function for modelling costs at scale

Alastair Ager Gerald Bloom Anna Vassall Institute of Global Health and Development, Queen Margaret University Institute of Development Studies, University of Sussex London School of Hygiene & Tropical Medicine Discussants Facilitator

Situating interventions in the policy context Chris Colvin University of Cape Town

Policy study main findings 1. Ownership: Many people felt that TB-IPC was seen as “everybody’s business and [therefore] nobody’s business”. 2. Advocacy: A general lack of TB activism, or any broader civil society mobilization to address TB concerns. 3. Evidence: A lack of new, clear and compelling effectiveness or cost-effectiveness evidence for TB-IPC interventions might make policy advocacy difficult 4. Implementation context: TB programmes were described as highly institutionalized and routinized and as stigmatized by HCWs (rigid professional hierarchies varied TB risk perceptions, persistent blind spot around where the real TB risk is located). 5. Key role of system level change: Success stories were only described at the clinic level, sub-district or district levels. TB-IPC is a problem that cannot be solved (only) by individuals or by national government.

Discussion Incorporating complexity into intervention design

Summary Chris Colvin Karina Kielmann University of Cape Town Queen Margaret University

Thanks & close Alison Grant London School of Hygiene & Tropical Medicine and Africa Health Research Institute

Our thanks to The nearly 7,000 people who participated: patients, clinic visitors, health care workers, policymakers and other government staff, engineers, architects, activists, researchers, Managers and staff at the 12 participating clinics Participants at System Dynamics and Patient Flow workshops Colleagues in Western Cape Department of Health: Andrew Boulle, Arne von Delft, Gavin Reagon, Bart Willems Our project oversight group: Thilo Govender, Graeme Meintjes, Vanessa Mudaly, Lindiwe Mvusi, Jacqui Ngozo, Helen Schneider Ethics committees at LSHTM, UCT, UKZN, and QMU

The extended team Investigators and core team: Kathy Baisley, Peter Beckwith, Adrienne Burrough, Chris Colvin, Karin Diaconu, Indira Govender, Alison Grant, Idriss Kallon, Aaron Karat, Karina Kielmann, Hayley MacGregor, Nicky McCreesh, Janet Seeley, Alison Swartz, Amy Thomas, Anna Vassall, Anna Voce, Richard White, Tom Yates, Gimenne Zwama University of KwaZulu-Natal: Bavashni Govender, Zama Khanyile, Nokuthula Lushaba, Nonhlanhla Maphumulo, Sharmila Rugbeer, Thandeka Smith Africa Health Research Institute: Nkosingiphile Buthelezi, Njabulo Dayi, Siphephelo Dlamini, Yutu Dlamini, Anita Edwards, Dickman Gareta, Patrick Gabela, Emmerencia Gumede, Sashin Harilall, Kobus Herbst, Mandla Khoza, Nozi Khumalo, Zilethile Khumalo, Nondumiso Kumalo, Richard Lessells, Sithembiso Luthuli, Sinethemba Mabuyakhulu, Nonhlanhla Madlopha, Thabile Mkhize, Duduzile Mkhwanazi, Zinhle Mkhwanazi, Zodwa Mkhwanazi, Sashen Moodley, Sihle Mthethwa, Xolile Mpofana, Sphiwe Mthethwa, Nozipho Mthethwa, Silindile Mthembu, Sanele Mthiyane, Vanisha Munsamy, Tevania Naidoo, Nompilo Ndlela, Thandekile Nene, Sabelo Ntuli, Nompumulelo Nyawo, The PIP CRAs, Anand Ramnanan, Aruna Sevakram, Sizwe Sikhakane, Zizile Sikhosana, Theresa Smit, Marlise Venter, Precious Zulu University of Cape Town: Amy Burdzik, Ruvimbo Chigwanda, Suzanne Key, Aphiwe Makalima, Godfrey Manuel, Phumzile Nywagi, Anathi Mngxekeza, Awethu Gawulekapa, Precious Mathenjwa, Samantha Moyo, Seonaid Kabiah, Sinead Murphy, Siphokazi Adonisi, Siphosethu Titise, Tamia Jansen, Yolanda Qeja Queen Margaret University: Jenny Falconer, Kitty Flynn, Claire McLellan London School of Hygiene & Tropical Medicine: Arminder Deol, Naomi Stewart

Funders

Further information is on our website: search for “Umoya omuhle” To access publications, please visit https://www.lshtm.ac.uk/research/centres-projects-groups/uo#publications These slides and the session recording will also be on the website