September 2019 A collection published by The Harvard Global Health Institute and The BMJ

Universal health coverage UNIVERSAL HEALTH COVERAGE

EDITORIAL 1 Delivering on the promise of universal health coverage Ashish Jha, Fiona Godlee, Kamran Abbasi

ANALYSIS 2 Rethinking assumptions about delivery of healthcare: implications for universal health coverage September 2019 EDITORIAL OFFICES Jishnu Das, Liana Woskie, Ruma Rajbhandari, Kamran Abbasi, Ashish Jha The Editor, The BMJ BMA House, Tavistock Square 7 Climate change threatens the achievement of effective universal healthcare London, UK, WC1H 9JR Email: [email protected] Renee N Salas, Ashish K Jha Tel: + 44 (0) 20 7387 4410 Fax: + 44 (0) 20 7383 6418 BMJ - Beijing 13 Achieving universal health coverage for mental disorders A1203 Tian Yuan Gang Center Vikram Patel, Shekhar Saxena East 3rd Ring North Road Chaoyang District 16 Motivating provision of high quality care: it is not all about the money Beijing 100027 China Mylène Lagarde, Luis Huicho, Irene Papanicolas Telephone: +86 (10) 5722 7209 BMJ - Hoboken BMJ Publishing Inc 21 Overcoming distrust to deliver universal health coverage: lessons from Ebola Two Hudson Place Liana R Woskie, Mosoka P Fallah Hoboken, NJ 07030 Tel: 1- 855-458-0579 email [email protected] 26 Global health security and universal health coverage: from a marriage of convenience BMJ - Mumbai 102, Navkar Chamber, A Wing to a strategic, effective partnership Marol, Andheri - Kurla Road Andheri (East) Mumbai 400059 Clare Wenham, Rebecca Katz, Charles Birungi, Lisa Boden, Mark Eccleston-Turner, Lawrence Gostin, Tel: +91 22-40260312/13/14 Renzo Guinto, Mark Hellowell, Kristine Husøy Onarheim, Joshua Hutton, Anuj Kapilashrami, Email: [email protected] Emily Mendenhall, Alexandra Phelan, Marlee Tichenor, Devi Sridhar BMJ - Noida Mindmill Corporate Tower 6th Floor, 24 A, Film City Sector 16 A ESSAY Noida 201301 Telephone: + 91 120 4345733 - 38 Email: [email protected] 32 How moves towards universal health coverage could encourage poor quality drugs BMJ - Singapore Suntec Tower Two Elizabeth Pisani 9 Temasek Boulevard, #29-01 Singapore 038989 Tel: +65 3157 1399 Email: [email protected] OPINION BMJ - Sydney Australia 35 Universal quality health coverage—committing to a healthier and more productive Telephone: +61 (0)2 8041 7646 Email: [email protected] ­society Twitter: Follow the editor, Fiona Godlee @fgodlee and The BMJ at twitter.com/bmj_latest Jeremy Veillard, Edward Kelley, Sepideh Bagheri Nejad, Francesca Colombo, Tim Evans, Niek Klazinga BMA Members’ Enquiries Email: [email protected],uk 37 Non-communicable diseases must be part of universal health coverage Tel: + 44 (0) 20 7383 6955 Advertising Sania Nishtar Email: [email protected] Tel: + 44 (0) 20 3655 5611 39 Building more effective health coverage in Argentina Reprints Email: [email protected] Adolfo Rubinstein Tel: + 44 (0) 7866 262 344 Subscriptions 41 Investing in community health workers accelerates universal health coverage Email: [email protected] Tel: + 44 (0) 20 7111 1105 Wilhelmina Jallah, Francis Kateh, Raj Panjabi Other resources Other contacts: http://www.bmj.com/about-bmj 43 Will artificial intelligence help universalise healthcare? Advice to authors: http://www.bmj.com/about-bmj/resources-authors To submit an article: submit.bmj.com Satchit Balsari The BMJ is published by BMJ Publishing Group Ltd, a wholly owned subsidiary of the British Medical Association. The BMA grants editorial freedom to the Editor of The BMJ. The views ­expressed in the journal are those of the authors and may not necessarily­ comply with BMJ policy. The BMJ follows guidelines on editorial ­independence produced by the World Association of Medical Editors (www.wame. org/wamestmt.htm#independence) and the code on good publication practice produced by the Committee on Publication Ethics (www.publicationethics.org.uk/guidelines/). The BMJ is intended for medical professionals and is provided without warranty, express or implied. Statements in the journal are the responsibility­ of their authors and advertisers and not authors’ ­institutions, the BMJ Publishing Group, or The BMJ unless otherwise specifi ed or determined by law. Acceptance of advertising does not imply endorsement. These articles are part of a series commissioned by The BMJ based on an idea from the Harvard Global Health Institute. The BMJ retained full To the fullest extent permitted by law, the BMJ Publishing Group shall not ­editorial control over external peer review, editing, and publication. Harvard Global Health Institute paid the open access fees. The article by be liable for any loss, injury, or damage resulting from the use of The BMJ Wenham et al is taken from BMJ Global Health. or any information in it whether based on contract, tort, or otherwise. Readers are advised to verify any information they choose to rely on. Indexing The BMJ @BMJ Publishing Group Ltd 2019 Please do not use the page numbers given in this edition when citing or linking to content in The BMJ. This printed Academic Edition is a monthly digest of selected content from The BMJ prepared primarily for an international and academic audience. Please be aware that The BMJ is an online All Rights Reserved. No part of this publication may be reproduced, stored journal, and the online version of the journal and each article at thebmj.com is the complete version. Please note that only the online article in a retrieval system, or transmitted in any form or by any other means, locator is required when indexing or citing content from The BMJ. We recommend that you use the Digital Object Identifier (doi) available online at electronic, mechanical, photocopying, recoding, or otherwise, without the top of every article and printed in each article in this edition for indexing. The citation format is given on each article. prior permission, in writing, of The BMJ. UNIVERSAL HEALTH COVERAGEFEATURE

Delivering on the promise of universal health coverage A new initiative to focus on improving healthcare delivery systems

niversal health coverage—the course, how is it possible to ensure that the know little about how best to deliver what is notion that people should coverage that is created is worth having? known to work. But that doesn’t mean that be able to access healthcare As Teerawattananon and colleagues have there aren’t generalisable principles, and services regardless of their pointed out, setting priorities early helps our hope is that The BMJ can be a vehicle for ability to pay, and do so with nations make better decisions about how best furthering our knowledge about how to do Ufinancial protection—is a major focus for G7 to achieve these goals, especially when faced universal health coverage well. policy makers and others around the globe. It with scarce resources.2 We are at a critical juncture in global health. is an important component of the sustainable Herein lies one of the largest challenges of The world has increasingly come to realise development goals, leading many countries to universal health coverage. The safety, quality, that we are interdependent and that a poor increase their use of scarce public resources and efficiency of most healthcare delivery performing health system in one place is a to ensure their citizens are covered. The systems (across high, middle, and low income threat to us all. We must pull together to help motivation, of course, is compelling. People countries) are far from the best they could nations develop their own healthcare systems should not develop or die from preventable be. In many countries, especially those with and achieve effective universal coverage in or treatable conditions because they are poor, scarce resources, the quality of the underlying ways that are consonant with their history, and treatment should not lead to financial healthcare delivery system is so poor that it is culture, and values. We can all learn from bankruptcy. unclear whether increasing access to services each other, and learn we must, because good Simply prioritising universal health will do more good or more harm. Conservative intentions are a start but they are insufficient. coverage, however, will not be enough to estimates put unsafe medical care as one Competing interests: We have read and understood achieve its main goals. Even if nations are of the top 10 causes of human harm, with BMJ policy on declaration of interests and have no able to identify the necessary resources, adverse medical events affecting 8-15% of relevant interests to declare. fundamental questions about achieving hospital inpatients in high income countries Provenance and peer review: Commissioned; not effective universal coverage remain and even more in low and middle income externally peer reviewed. 3 unanswered. Without these answers, simply countries. In some places, patients who visit Ashish Jha, director1 pushing for universal coverage may waste physicians are more likely to get the wrong Fiona Godlee, editor in chief2 precious resources without achieving the diagnosis and harmful treatment than they Kamran Abbasi, executive editor2 4 5 important goals of improved health and are to get the right ones. 1Harvard Global Health Institute, Cambridge, MA wellbeing of citizens in a way that is efficient While we strongly support the idea of 02138, USA for people and countries. universal health coverage, we do not know 2The BMJ, London, UK Although the primary focus is coverage how to ensure we get the most for our Correspondence to: A Jha (financing healthcare services, ensuring investments. We do know, however, that given [email protected] adequate staff, etc), success depends on the millions of deaths that occur from poor the underlying healthcare delivery system. quality care, one way to improve the value of It also requires clarity about the trade-offs. that investment is to improve the quality of the 1 Massachusetts Budget and Policy Center. Governments have limited resources but underlying delivery system. Massachusetts state budget. 2016. http://massbudget. many important priorities. What will not To this end, the Harvard Global Health org/browser/index.php. be funded because governments focus on Institute and The BMJ have produced a 2 Teerawattananon Y, Luz A, Kanchanachitra C, Tantivess S; Prince Mahidol Award Conference Secretariat. Role universal coverage? Even in Massachusetts, collection of articles on effective universal of priority setting in implementing universal health one of the wealthier US states, investments coverage. We agree with the underlying coverage. BMJ 2016;532:i244. PubMed doi:10.1136/ in achieving universal health coverage for its goals—that everyone around the world bmj.i244 population of seven million people coincided deserves access to healthcare services 3 Jha AK, Larizgoitia I, Audera-Lopez C, Prasopa-Plaizier N, Waters H, Bates DW. The global burden of unsafe with substantial reductions in spending on when they are ill and encounters with the medical care: analytic modelling of observational public health, education, and infrastructure.1 healthcare system should make people better studies. BMJ Qual Saf 2013;22:809-15. PubMed If policy makers decide that universal without bankrupting them. But there is a vast doi:10.1136/bmjqs-2012-001748 4 Das J, Hammer J, Leonard K. The quality of health coverage is worth supporting, they will gap between those two goals and our ability medical advice in low-income countries. J Econ need better information on how to implement to deliver them. That vast gap is mainly due to Perspect 2008;22:93-114. PubMed doi:10.1257/ it. How should investments be structured? deficiencies in knowledge about both optimal jep.22.2.93 5 Das J, Holla A, Das V, Mohanan M, Tabak D, Chan B. Should it all be public funding or should approaches to financing and effective models In urban and rural India, a standardized patient study governments fund private insurance and for healthcare delivery. So little is known showed low levels of provider training and huge private providers? How should governments about how to do this well—partly because quality gaps. Health Aff (Millwood) 2012;31:2774-84. ensure that there is a legal and regulatory each nation is unique with a different set of PubMed doi:10.1377/hlthaff.2011.1356 framework to manage the legal rights that needs and a different path to achieving true, Cite this as: BMJ 2016;353:i2216 such programmes might confer? And, of effective universal health coverage. We also http://dx.doi.org/10.1136/bmj.i2216

the bmj | BMJ 2016;353:i2216 | doi: 10.1136/bmj.i2216 1 UNIVERSAL HEALTH COVERAGE

Rethinking assumptions about delivery of healthcare: implications for universal health coverage Simply providing more resources for universal coverage is not enough to improve health, argue Jishnu Das and colleagues. We also need to ensure good quality of care

e are at an inflection point has made UHC his top priority for the of necessary services is foundational to in global health. People agency. the performance of health systems; a are living longer, healthier UHC can achieve its primary objective of system that cannot accurately diagnose lives than ever before, and creating better health, but to do so, patients or manage patients will not deliver the we are rightly celebrat- must have access to services that are high improved health outcomes implicit in the Wing disease focused programmes that have quality. This idea of “effective UHC” is UHC agenda. greatly reduced or eradicated diseases such not new. It has long been recognised that as smallpox and river blindness. Better diag- translating healthcare into health outcomes Assessing the evidence and identifying the nosis and treatment of HIV/AIDS, malaria, requires that services meet some basic problems and other diseases have saved countless standard of quality.3 However, without Our synthesis relies on recent studies of the lives.1 2 Yet, as populations age and the systematic data on quality, the working quality of clinical practice and its determi- burden of morbidity grows more complex, assumption has been that adequately nants in low and middle income countries the limitations of programmes focused on trained doctors and nurses with access (LMICs). In the absence of administrative single ­diseases have become increasingly to infrastructure (such as well equipped data sources or information from patient evident. facilities and medicines) will be sufficient charts (which are rare or of doubtful quality Policy makers have shifted towards a to guarantee adequate quality. Emerging in many of these countries), these studies broader “systems” view of universal health data suggest that this understanding may have used surveys of healthcare provid- coverage (UHC)—one that seeks to provide be incorrect. For example, even when ers (medical vignettes and standardised all people with access to essential health resources are in place in countries as far patients) to measure two related but sepa- services without financial hardship—as the afield as Bangladesh and Uganda, health rate things: what providers know about defining approach to improve the health of systems are unable to ensure that doctors managing common medical conditions the world’s poorest people. As one of the show up to work, with absence rates and how they actually practise in clinical key focuses of the sustainable development ranging from 40% to 60%.4 5 And when settings (see appendix on bmj.com). Three goals, UHC has become a rallying principle they do, the services they provide are far key issues emerge from this evidence and for all countries. Indeed, the new director below any acceptable standard. are discussed below. general of the World Health Organization We focus on one aspect of quality— effectiveness, or the degree to which Without quality, access may be irrelevant patients receive timely and accurate Health policy efforts often invest substan- KEY MESSAGES diagnoses and evidence based treatments tially in programmes that have the primary 6 • Availability of health advisers is not for their conditions. Other domains objective of increasing the use of healthcare the main problem in most countries of quality, such as patient safety and services, such as the number of treatment 8 • Simply providing access to trained patient centredness (table 1), are equally episodes or health visits per patient. But medical staff and facilities does not important. However, the effective provision emerging data suggest that this focus on guarantee universal access to quality care Table 1 | Essential elements of quality healthcare (adapted from Scott and Jha7) • A weak link between medical quali- Domain Subcategorisation Example measures fications and medical knowledge Safety Adverse events—eg, due to medical devices or Rate of prescriptions above the maximum daily dose implies that providers without any medicines, including substandard and falsified Rate of infection or foreign objects left during surgical formal medical training can provide medicines procedures higher quality care than fully trained Healthcare acquired conditions Cases of hospital acquired pneumonia among doctors inpatients Effectiveness Timely and accurate diagnosis Rate of correct diagnosis of cervical cancer In many countries large gaps exist • Evidence based treatment, including Rate of appropriate treatment for patients presenting between what doctors know and what appropriate follow-up and management with childhood diarrhoea they actually do Rate of glycaemic control among patients diagnosed • New approaches are needed to tackle with diabetes systems that produce medical profes- Patient Patient experience Rate of patients who would recommend their provider centredness to a family or friend sionals who are poorly trained, under- Patient reported outcomes Patients reporting adequate or high functional status motivated, and underused after surgery

2 doi: 10.1136/bmj.k1716 | BMJ 2018;361:k1716 | the bmj UNIVERSAL HEALTH COVERAGE

Table 2 | Key findings of studies using standardised patients13-17 No of healthcare % Correctly managed or referred No of standardised visits / practitioners % With correct No unnecessary Some unnecessary % given unnecessary Location of study Conditions studied patients included diagnosis drugs given drugs given antibiotics India: Delhi (urban) Angina, asthma, and diarrhoea 17 250 23 46 NA NA Tuberculosis 17 250 NA 8 21 54 Madhya Pradesh (rural) Angina, asthma and diarrhoea 22 677 12* 8 36 35 Bihar (rural) Childhood diarrhoea NA 340 3 0 17 NA Childhood pneumonia NA 340 8 14 60 NA China: Shaanxi Province (rural) Dysentery and angina 4 82 37 24 52 NA Sichuan, Shaanxi, and Tuberculosis 4 138 15 25 40 51 Anhui Provinces (rural) Kenya: Nairobi (urban) Angina, asthma, diarrhoea, 14 166 32* 22 53 55 and tuberculosis *Denominators for denoted percentages are limited to cases in which a diagnosis was given. All other rates have a denominator of overall cases. The proportion of presentations that received a diagnosis ranged from 6% in the Bihar childhood diarrhoea case to 90% in the China tuberculosis case. getting people in the door may not lead to treatment. This method facilitates a “blind providers who see too many patients and improved health. audit” since the same clinical cases can be do not have the time to carefully evaluate We often begin with the assumption that presented to providers with a wide range of or manage them may be incorrect. Clinical a key feature of many health systems in training and qualifications.13 14 observation studies show that most LMICs is the lack of access to healthcare In India, China, and Kenya most cases primary care providers see too few patients, services. We measure access by counting were incorrectly diagnosed, and, even rather than too many (fig 1). The average the number and proximity of formal using a very lenient definition, simple healthcare provider working in a public healthcare providers who work in official medical conditions were correctly clinic in rural India, who provides services clinics. In reality, in many countries, people managed a minority of the time. that are free at the point of use, sees 5.7 may have access to multiple healthcare Although standardised patients in Kenya patients a day, spending only three minutes providers with varying qualifications generally received higher quality care with each. In Tanzania, Senegal, and rural and connections to the formal healthcare than those in India and China, 90% of Madhya Pradesh (India), doctors in public sector. The average village in rural Madhya angina presentations in Nairobi were still primary health clinics spend a mere 30 to Pradesh—one of the poorest states in misdiagnosed as pneumonia.14 Across the 40 minutes a day seeing patients. India—has 11 healthcare providers within board, studies find frequent misdiagnosis, 3 km of the village,9 most of whom have overuse of antibiotics and other drugs, and Qualifications do not equal clinical knowledge no formal training.10 However, informal underuse of inexpensive but potentially Poor quality is often assumed to be due to providers are often not counted when lifesaving diagnostics and therapies in both the large number of informal (ie, untrained) assessing key measures of access such as public and private sector clinics; all have the ratio of clinicians to patients. serious repercussions for health outcomes . In other countries, non-physician and expenditure. clinicians are an integral and sizeable Poor quality is not unique to primary . part of the state machinery but are often care. Another stark example is institutional . excluded when assessing in human childbirth. Incentive schemes to encourage . 11 12 resources. Studies that count all women to deliver in public facilities Average daily time .

providers show that access to healthcare increased the number of institutional seeing patients (hours) . is often better than historically imagined deliveries in countries such as Malawi,18 in low resource settings. Official statistics India,19 20 and Rwanda21 but did not that focus only on formal physicians per improve child or maternal outcomes. Why population miss this important point. not? It is not for the lack of availability of Birbhum, India Vietnam district Given that access, more leniently infrastructure and medicines. According (public clinics) (private clinics) Vietnam commune defined, is less of a problem, where do the to WHO surveys, lifesaving treatments for Madhya Pradesh,Madhya India Pradesh, India challenges lie? Primarily, it is the quality of women giving birth are widely available care that patients receive when they access and used in most health facilities across Fig 1 | Total time spent by healthcare healthcare providers. Table 2 summarises countries. However, the availability of providers with patients over a day. The the results of studies that use standardised these essential treatments is not associated sample from Madhya Pradesh, includes 199 patients—people recruited from local with better maternal outcomes.22 Poor private providers (mostly untrained) and communities and extensively trained to implementation, delays in diagnosis and 119 providers in public clinics. The sample from Birbhum, is 256 providers in rural present the same set of standard symptoms treatment, and silos of care have been locations, most of whom are not formally to multiple providers—to assess quality. hypothesised to at least partly explain trained. The survey from Vietnam is based The standardised patients presented with excessive mortality and morbidity. on a representative sample of 214 commune simple clinical conditions to ensure no Finally, the hypothesis that poor quality health facilities (similar to primary health disagreement on the correct diagnosis or may be due to overwhelmed primary care centres) and 171 district hospitals23 15 24 the bmj | BMJ 2018;361:k1716 | doi: 10.1136/bmj.k1716 3 UNIVERSAL HEALTH COVERAGE

Adherence to checklist for for multiple conditions presented to does in Delhi’s public sector or Vietnam’s median Kenyan nurse doctors through medical vignettes in five district hospitals, medical knowledge often Medical ocer Nurse sub-Saharan African countries. Although does not translate into high quality clinical  fully trained doctors are more likely interactions.28 A recent systematic review  than nurses to know what questions of consultation time, our best measure of to ask and examinations to perform, effort, across 68 countries and 28 million  Standard deviaton there is considerable overlap between consultations found that the average con- - the distributions (fig 2). Within every sultation “varied from 48 seconds in Bang- country, the top 20-50% of nurses are ladesh to 22.5 minutes in Sweden.” In most - more knowledgeable than the poorest countries, consultation times averaged less performing 25% of doctors. Even between than 10 minutes, and in 15 countries less formally trained versus informally than 5 minutes.29 Short consultation times Kenya (n= ) Uganda (n= ) Nigeria (n=)Tanzania (n=) trained doctors, doctors with more formal were more prevalent in low income coun- Madagascar (n=) education may only modestly outperform tries, even in contexts where doctors were Fig 2 | Variations in medical knowledge of their informally educated peers (fig 3). seeing just a few patients a day.23 medical officers (fully trained doctors) and The translation of qualifications to Short consultation times imply that nurses assessed by World Bank’s Service knowledge varies across countries. The even when doctors know what to do, they Delivery Indicators Survey. The boxes show mean Kenyan nurse is more knowledgeable often fail to do it. There is a persistent, 25th percentile, median, and 75th percentile than 21% of doctors in Kenya, 78% of often sizeable, gap between what adherence to condition specific checklist doctors in Madagascar, 32% in Nigeria, providers say they will do when faced items for the common illnesses, with the whiskers giving the 10th and 90th percentiles 25% in Tanzania, and 63% in Uganda (fig with a hypothetical patient and what they 2). There are also wide differences across actually do when they see such a patient states in India: informal providers in high (fig 4). Emerging evidence finds large providers. However, even fully trained pro- performing states like Tamil Nadu are more “know-do” gaps in countries as diverse as viders with adequate access to infrastruc- knowledgeable than fully trained doctors Rwanda,31 Tanzania,32 India,28 China,30 ture often fail to deliver high quality care. in low performing states like Bihar. The and Vietnam.23 This know-do gap can be so This weak link between qualifications and link between qualifications (training) large that the providers without any formal quality reflects two related but conceptu- and medical knowledge is surprisingly medical training can provide higher quality ally separate issues. Firstly, the quality of weak. It is therefore wrong to assume that care than fully trained doctors.28 medical training varies considerably in populations with access to a fully trained The idea that the medical profession many countries. Tests of medical knowl- doctor in Madagascar enjoy better care than “has special knowledge … and will edge among physicians and non-physician 33 25 23 26 populations with access to a fully trained self-regulate” has already been clinicians in India, Vietnam, Nigeria, 34 35 27 nurse in Kenya. questioned. We are learning that Eastern Europe, and several countries doctors are humans who operate within in sub-Saharan Africa consistently show Clinical knowledge often fails to translate into complex systems. Because they respond large variations in within country knowl- clinical practice to incentives, the same doctors seem to edge, with sizeable numbers of untrained, Medical knowledge is only loosely tied to provide more effort (and deliver higher non-physician clinicians who are more actual clinical practice. Providing high quality care) in private clinics than knowledgeable than their fully trained quality clinical care requires both knowl- in public ones, even when structural counterparts. edge and effort, and when the average resources are held equal. In a Beijing Figure 2 documents adherence to a clinical interaction lasts 90 seconds, as it hospital, when standardised patients medically necessary checklist of questions presenting with viral pharyngitis told about medical history and examinations doctors they would purchase medicines Standardised patients Clinical vignette from an external pharmacist (rather Intermediate, undergraduate, or bachelor (n=)  Specialist or masters (n=) than the hospital pharmacy from which   the prescriber receives a salary bonus), antibiotic prescriptions fell from 77% to 36   11%. This 66 percentage point difference suggests doctors knew that prescribing No of provider s

correct ttreatment  Providers ordering antibiotics was unhelpful but were swayed   by financial incentives. Bihar China Delhi Potential solutions     Diarrhoea Tuberculosis We have focused on just one component Knowledge score (ORS) (AFB or CXR) of quality: effectiveness. Understanding Fig 3 | Relation between medical qualification Fig 4 | Differences between how providers whether similar patterns arise for safety and knowledge among doctors in Vietnam, said they would manage diarrhoea and and patient centred care is critical, although as assessed by medical vignettes. The circles turberculosis in clinical vignettes and what there is little reason to believe it would not. show the number of providers in each bin they actually did with standardised patients The data come from only around a dozen of 0.1 standard deviation across the entire presenting with the symptoms in the countries, but they include India and China, distribution. The corresponding density vignettes (ORS=oral replacement solution, where a large proportion of the world’s plots (relative scale) are calculated from the AFB=acid fast bacilli test, CXR=chest underlying unbinned distributions23 radiography) 13 17 30 poorest people live. Although standardised

4 doi: 10.1136/bmj.k1716 | BMJ 2018;361:k1716 | the bmj UNIVERSAL HEALTH COVERAGE

patients cannot fully capture all clinical sce- Conclusion narios (for practical and ethical reasons), Task shifting and new approaches to deliv- the data that have emerged in recent years ery are just two examples of the kind of This is an Open Access article distributed in suggest the same patterns: big quality prob- innovation needed to achieve effective accordance with the Creative Commons Attribution lems, a weak link between qualifications and UHC. Reaching the goals of UHC requires Non Commercial (CC BY-NC 4.0) license, which knowledge, and a large gap between knowl- not just more money, but better money. We permits others to distribute, remix, adapt, build upon this work non-commercially, and license edge and practice. The evidence suggests need additional research and policy work their derivative works on different terms, provided that countries need to incorporate quality that questions baseline assumptions and the original work is properly cited and the use is into their UHC plans at an early stage. normative, or prescriptive, frameworks. We non-commercial. See: http://creativecommons.org/ Whether efforts to achieve UHC will must understand the world as it is, not as licenses/by-nc/4.0/. translate into better health outcomes we wish it to be. Healthcare providers may depends on how we execute these efforts, make errors, but they often make the same and this in turn will determine whether errors again and again, and therefore make 1 Bendavid E, Holmes CB, Bhattacharya J, Miller G. we are able to move from a simple access “predictable” mistakes; these mistakes are HIV development assistance and adult mortality oriented definition of UHC to truly effective indicative of a broken system. If this pre- in Africa. JAMA 2012;307:2060-7. doi:10.1001/ UHC. Emerging data challenge models of dictability is recognised and modelled in jama.2012.2001 2 World Health Organization. Eliminating river care that assume that qualified providers policies and strategies to improve global blindness: highlights from TDR’s making a difference. in well resourced clinics guarantee quality. health, we can make important advances. 30 years of research and capacity building in tropical New approaches are needed to ensure that Such recognition has the potential to trans- diseases. WHO, 2007. broader investments in healthcare actually form how healthcare is delivered in low 3 Spector JM, Agrawal P, Kodkany B, et al. Improving quality of care for maternal and newborn health: lead to better health outcomes, especially income contexts, ultimately improving the prospective pilot study of the WHO safe childbirth for poorer people. lives of billions. checklist program. PLoS One 2012;7:e35151. New approaches need to tackle systems doi:10.1371/journal.pone.0035151 Additional material available on bmj.com: 4 Chaudhury N, Hammer JS. Ghost doctors: that produce medical professionals who are Methods for collecting data on effectiveness of absenteeism in Bangladeshi health facilities. World poorly trained, undermotivated, and often healthcare Bank, 2004. https://openknowledge.worldbank.org/ handle/10986/17167 assigned to clinics with no peers or mentors Contributors and sources: All authors contributed to the design of the manuscript as well as interpretation 5 Chaudhury N, Hammer J, Kremer M, and insufficient patient volume to hone Muralidharan K, Rogers FH. Missing in action: skills. These providers consequently leave of the data. JD provided data for figures used throughout the piece. JD, LW, and RR drafted the teacher and health worker absence in developing many patients, particularly those with few manuscript. All authors then reviewed and provided countries. J Econ Perspect 2006;20:91-116. resources, receiving care that is unhelpful comment. JD, LW and AJ received comments and doi:10.1257/089533006776526058 6 Institute of Medicine, Committee on Quality of and often harmful. updated the manuscript, after which all authors reviewed and provided feedback and, finally, approval. Health Care in America. Crossing the quality chasm: This will not be an easy process. But clear Mary Dixon Woods also provided substantive feedback a new health system for the 21st century. National examples are emerging where these efforts throughout the development of the piece. The findings, Academies Press, 2001. interpretations, and conclusions expressed here are 7 Scott KW, Jha AK. Putting quality on the global health are bearing fruit: mid-level providers who agenda. N Engl J Med 2014;371:3-5. doi:10.1056/ provide high quality care, whether they are those of the authors and do not necessarily represent the views of the World Bank, its executive directors, NEJMp1402157 anaesthesia assistants in rural Nepalese or the governments they represent. Rifat Atun, Niek 8 Institute of Medicine, Committee to Design a hospitals or nurses managing HIV care in Klazinga, Fiona Godlee, Ed Kelley, Sheila Leatherman, Strategy for Quality Review and Assurance in Medicare. InLohr KN, ed. Medicare: A strategy large parts of Africa.37-39 Initiatives to tackle and Sujatha Rao were part of the working group for the series. for quality assurance. National Academies Press, the availability of doctors in rural areas 1990. can focus on non-physician providers and Competing interests: We have read and understood 9 Das J, Mohpal A. Socioeconomic status and BMJ policy on declaration of interests and have no quality of care in rural India: new evidence from training them to be as good, if not better, at relevant interests to declare. provider and household surveys. Health Aff 40 41 providing certain types of care. (Millwood) 2016;35:1764-73. doi:10.1377/ Provenance and peer review: Commissioned; hlthaff.2016.0558 Similarly, countries are realising that externally peer reviewed. placing doctors in rural areas may mean 10 Banerjee A, Deaton A, Duflo E. Health, This article is part of a series commissioned by The health care, and economic development: that they see only few patients a day. An BMJ based on an idea from the Harvard Global Health wealth, health, and health services in rural alternative is to bring patients from rural Institute. The BMJ retained full editorial control over Rajasthan. Am Econ Rev 2004;94:326-30. areas to urban centres with better facilities, external peer review, editing, and publication. Harvard doi:10.1257/0002828041301902 Global Health Institute paid the open access fees. 11 Sudhinaraset M, Ingram M, Lofthouse HK, Montagu D. as has been tried with considerable success What is the role of informal healthcare providers Jishnu Das, lead economist1 using ambulance systems in India and in developing countries? A systematic review. 2 medical buses in Brazil.42 43 Liana Woskie, assistant director PLoS One 2013;8:e54978. doi:10.1371/journal. Ruma Rajbhandari, instructor3 pone.0054978 Unfortunately, there are other systematic 12 Mullan F, Frehywot S. Non-physician clinicians Kamran Abbasi, executive editor4 design problems where our knowledge in 47 sub-Saharan African countries. 5 base remains low. For instance, evidence Ashish Jha, professor of international health Lancet 2007;370:2158-63. doi:10.1016/S0140- 1 shows that when diagnosis and treatment World Bank, Washington, DC, USA 6736(07)60785-5 2 13 Das J, Kwan A, Daniels B, et al. Use of standardised are “bundled” so that healthcare providers Harvard Initiative on Global Health Quality, Cambridge, MA 02138, USA patients to assess quality of tuberculosis care: can earn higher incomes by ordering tests a pilot, cross-sectional study. Lancet Infect 3Harvard Medical School, Boston, MA, USA or prescribing drugs, their tendency to do Dis 2015;15:1305-13. doi:10.1016/S1473- 4The BMJ, London, UK 3099(15)00077-8 so increases.36 Breaking the link between 5Department of Health Policy and Management, Harvard 14 Daniels B, Dolinger A, Bedoya G, et al. Use diagnoses, drug sales, and laboratory tests T H Chan School of Public Health, Harvard Global Health of standardised patients to assess quality of can reduce unnecessary tests and drug Institute, Boston healthcare in Nairobi, Kenya: a pilot, cross- sectional study with international comparisons. usage. How to do so in an efficient manner, Correspondence to: A K Jha BMJ Glob Health 2017;2:e000333. doi:10.1136/ however, remains an open question. [email protected] bmjgh-2017-000333 the bmj | BMJ 2018;361:k1716 | doi: 10.1136/bmj.k1716 5 UNIVERSAL HEALTH COVERAGE

15 Das J, Holla A, Das V, Mohanan M, Tabak D, Chan B. 25 Das J, Hammer J. Which doctor? Combining vignettes 36 Currie J, Lin W, Meng J. Addressing antibiotic In urban and rural India, a standardized patient study and item response to measure clinical competence. abuse in China: an experimental audit study. showed low levels of provider training and huge J Dev Econ 2005;78:348-83. doi:10.1016/j. J Dev Econ 2014;110:39-51. doi:10.1016/j. quality gaps. Health Aff (Millwood) 2012;31: jdeveco.2004.11.004 . jdeveco.2014.05.006 2774-84. doi:10.1377/hlthaff.2011.1356 26 Wane W, Martin GH. Education and health services 37 Shumbusho F, van Griensven J, Lowrance D, 16 Sylvia S, Shi Y, Xue H, et al. Survey using incognito in Uganda: data for results and accountability. World et al. Task shifting for scale-up of HIV care: standardized patients shows poor quality care in Bank, 2013. evaluation of nurse-centered antiretroviral China’s rural clinics. Health Policy Plan 2015;30: 27 Peabody JW, DeMaria L, Nguyen SN, Smith O, treatment at rural health centers in Rwanda. PLoS 322-33. doi:10.1093/heapol/czu014 Hoth A. Quality of care in six eastern European countries Med 2009;6:e1000163. doi:10.1371/journal. 17 Mohanan M, Vera-Hernández M, Das V, et al. using clinical performance and value vignettes: a cross- pmed.1000163 The know-do gap in quality of health care for sectional study. Bull World Health Organ 2017. 38 Cohen R, Lynch S, Bygrave H, et al. Antiretroviral childhood diarrhea and pneumonia in rural India. 28 Das J, Hammer J. Money for nothing: the dire treatment outcomes from a nurse-driven, JAMA Pediatr 2015;169:349-57. doi:10.1001/ straits of medical practice in Delhi. World Bank, community-supported HIV/AIDS treatment jamapediatrics.2014.3445 200510.1596/1813-9450-3669 . programme in rural Lesotho: observational cohort 18 Godlonton S, Okeke EN. Does a ban on informal 29 Irving G, Neves AL, Dambha-Miller H, et al. assessment at two years. J Int AIDS Soc 2009;12:23. health providers save lives? Evidence from Malawi. International variations in primary care physician doi:10.1186/1758-2652-12-23 J Dev Econ 2016;118:112-32. doi:10.1016/j. consultation time: a systematic review of 39 Mavalankar D, Sriram V. Provision of anaesthesia jdeveco.2015.09.001 67 countries. BMJ Open 2017;7:e017902. services for emergency obstetric care through 19 Powell-Jackson T, Mazumdar S, Mills A. Financial doi:10.1136/bmjopen-2017-017902 task shifting in South Asia. Reprod Health incentives in health: New evidence from India’s 30 Sylvia S, Xue H, Zhou C, et al. Tuberculosis detection Matters 2009;17:21-31. doi:10.1016/S0968- Janani Suraksha Yojana. J Health Econ 2015;43: and the challenges of integrated care in rural China: 8080(09)33433-3 154-69. doi:10.1016/j.jhealeco.2015.07.001 A cross-sectional standardized patient study. PLoS 40 Sanne I, Orrell C, Fox MP, et al, CIPRA-SA Study 20 Mohanan M, Bauhoff S, La Forgia G, Babiarz KS, Med 2017;14:e1002405. doi:10.1371/journal. Team. Nurse versus doctor management of HIV- Singh K, Miller G. Effect of Chiranjeevi Yojana on pmed.1002405 infected patients receiving antiretroviral therapy institutional deliveries and neonatal and maternal 31 Gertler P, Vermeersch C. Using performance (CIPRA-SA): a randomised non-inferiority trial. outcomes in Gujarat, India: a difference-in-differences incentives to improve health outcomes. World Bank, Lancet 2010;376:33-40. doi:10.1016/S0140- analysis. Bull World Health Organ 2014;92:187-94. 201210.1596/1813-9450-6100 . 6736(10)60894-X doi:10.2471/BLT.13.124644 32 Leonard KL, Masatu MC. The use of direct clinician 41 Fairall L, Bachmann MO, Lombard C, et al. Task 21 Chari A, Okeke E. Can institutional deliveries reduce observation and vignettes for health services shifting of antiretroviral treatment from doctors newborn mortality? Evidence from Rwanda. RAND, quality evaluation in developing countries. Soc to primary-care nurses in South Africa (STRETCH): 2014. Sci Med 2005;61:1944-51. doi:10.1016/j. a pragmatic, parallel, cluster-randomised trial. 22 Souza JP, Gülmezoglu AM, Vogel J, et al. Moving socscimed.2005.03.043 Lancet 2012;380:889-98. doi:10.1016/S0140- beyond essential interventions for reduction of 33 Berwick D. The 9 changes needed to make health 6736(12)60730-2 maternal mortality (the WHO Multicountry Survey care more moral. 2016. https://www.advisory.com/ 42 Babiarz KS, Mahadevan SV, Divi N, Miller G. on Maternal and Newborn Health): a cross-sectional daily-briefing/2016/04/12/berwick. Ambulance service associated with reduced study. Lancet 2013;381:1747-55. doi:10.1016/ 34 Brennan TA, Leape LL, Laird NM, et al. Incidence probabilities of neonatal and infant mortality in two S0140-6736(13)60686-8 of adverse events and negligence in hospitalized Indian states. Health Aff (Millwood) 2016;35:1774- 23 World Bank. Quality and equity in basic health care patients. Results of the Harvard Medical 82. doi:10.1377/hlthaff.2016.0564 services in Vietnam: findings from the 2015 Vietnam Practice Study I. N Engl J Med 1991;324:370-6. 43 Marques AJdS. Lima MdS. O sistema estadual de district and commune health facility survey. World doi:10.1056/NEJM199102073240604 transporte em saúde de Minas Gerais: relato de Bank, 2016. 35 Landrigan CP, Parry GJ, Bones CB, Hackbarth AD, experiência. Revista de Administração Hospitalar e 24 Das J, Chowdhury A, Hussam R, Banerjee AV. The Goldmann DA, Sharek PJ. Temporal trends in rates Inovação em Saúde 2012;8:81-4. impact of training informal health care providers of patient harm resulting from medical care. N in India: A randomized controlled trial. Science Engl J Med 2010;363:2124-34. doi:10.1056/ Cite this as: BMJ 2018;361:k1716 2016;354:aaf7384. doi:10.1126/science.aaf7384 NEJMsa1004404 http://dx.doi.org/10.1136/bmj.k1716

6 doi: 10.1136/bmj.k1716 | BMJ 2018;361:k1716 | the bmj UNIVERSAL HEALTH COVERAGE

Climate change threatens the achievement of effective universal healthcare Minimising the health harms of climate change and optimising universal health coverage will only be achieved through an integrated agenda and aligned solutions, say Renee Salas and Ashish Jha

he sustainable development health outcomes. To track progress on SDG Changes in disease burdens (type and goals (SDGs) target many dif- 3.8, the World Health Organization (WHO) distribution) ferent aspects of human well- and World Bank created a service coverage The effects of climate change will interact being; they are interconnected index to measure the extent of coverage with other forces that affect health (box 1). and some might seem to create of “essential health services.” Although Non-communicable diseases accounted for Ttension (such as economic growth in SDG coverage has increased globally, only 22 71% of global deaths in 2016, and three 8 and ecological stewardship in SDGs 12 countries currently have a “high coverage” of the top causes (cardiovascular disease, and 15).1 These interconnections are par- index.2 chronic respiratory disease, and diabetes) ticularly clear for universal health coverage Climate change is already threatening are exacerbated by climate change, as is (UHC) (SDG 3.8), which will be substan- many health achievements of the past mental health.5 9-11 Climate change is also tially harder to achieve without climate 50 years and will continue to do so at an increasing the frequency and geographic action (SDG 13). Climate change threatens accelerated pace unless we take action.3 spread of infectious diseases. the very tenets of UHC; the regions of the WHO estimates that climate change will These disease burdens related to climate world most vulnerable to climate change cause an additional 250 000 deaths a year change pose added obstacles to UHC face the greatest difficulties in achieving it. by 2030, when taking into account just by increasing overall use and costs of United Nations countries agreed to five exposure pathways (undernutrition, healthcare.12 As UHC programmes seek achieve UHC by 2030, which requires malaria, diarrheal disease, dengue, and to define the essential services that they optimal access to essential, high heat).4 will cover and to build financial models quality services without sacrificing Our understanding of how climate for their costs, these growing and novel affordability.2 This extends beyond merely change affects health is still growing, but burdens will make appropriate coverage providing “coverage” and has three main we know it will have multiple direct and more challenging. In addition, tools used components: a broad set of healthcare indirect negative effects, including greater by policy makers for prioritisation in services must be accessible, affordable, heat related morbidity, undernutrition, coverage decisions will need to be updated and of sufficiently high quality to improve increasing water and foodborne illnesses, to reflect shifts in disease burdens from and mental health problems.5 The largest climate change. KEY MESSAGES driver for greenhouse gases globally is the combustion of fossil fuels,6 7 and Population displacement and migration • Climate change is threatening to the resultant air pollution leads to an The number of displaced people is pre- undermine the achievement of univer- additional seven million deaths annually.6 dicted to be 143 million by 2050 in just sal health care (UHC) through nega- Although the health effects of climate three regions (Latin America, sub-Saharan tive health outcomes and healthcare Africa, and South Asia), in part because change are wide reaching, they can still be 21 system disruptions mitigated if we take action now. of climate change. Displacement might • Climate change and UHC agendas Beyond direct health effects, climate be driven by property loss, resource short- bolster each other as they both strive change will make it more difficult to achieve ages, and conflicts. These consequences of to improve health and achieve health UHC. The global community has the climate change occur on the backdrop of equity urgent opportunity to tackle two pressing broader political and societal issues, such as immigration policies and conflict, show- Many regions of the world with challenges of our time: UHC and climate • ing the complexities of the problem.22 the highest vulnerability to climate change. In this piece, we discuss the Ensuring that a largely stationary change are also those with the lowest pathways through which climate change population can access a broad set of UHC coverage. These regions stand to will create barriers for achieving UHC and high quality services is hard enough; have enormous gains through an inte- how policy makers should mitigate these delivering UHC to a migratory population grated approach harms. is substantially more challenging. Chad, UHC plans should work to improve • How climate change threatens UHC for example, is experiencing increased the understanding of climate change, Achieving effective UHC even in the migration secondary to drought with use novel climate sensitive financial absence of climate change is difficult.8 concerns for strain on public health services frameworks, and incorporate the miti- 23 Climate change is a “meta problem,” cre- and health complications. Displaced gation of greenhouse gases ating strong headwinds that will make populations have distinct health related They should strive for evidence based • ensuring access to affordable, high qual- needs, as they may have different rates of climate adaptation that protects ity care more challenging (fig 1). Climate conditions, face mental health problems, health and prioritise health system change threatens UHC through five key and bring novel diseases. The influx of climate resiliency pathways. people alone might pose a challenge to the bmj | BMJ 2019;366:l5302 | doi: 10.1136/bmj.l5302 7 UNIVERSAL HEALTH COVERAGE

Climate change Threatens: Universal health coverage

Higher greenhouse gas concentrations Changes in disease burdens Accessibility

Population displacement and migration

Rising temperature and heat waves Rising poverty Quality

Intensi cation of extreme weather Disruption of healthcare infrastructue and care delivery

Higher sea levels Health workforce disruption and impairment Affordability

Fig 1 | How climate change threatens the achievement of universal health coverage

local healthcare systems—particularly in Disruption of healthcare infrastructure and provider, further exacerbating affordability locations with no or low coverage—as they care delivery concerns. struggle to manage the increased patient Extreme weather events related to climate volume and provide culturally sensitive change, like more intense hurricanes Health workforce disruptions and care. and floods, can cause structural damage impairments or power outages at healthcare facilities There is already a shortage and maldistri- Rising poverty (box 2). Even undamaged facilities can be bution of well trained healthcare workers 30 The World Bank estimates that climate affected by supply chain disruptions—due around the world. This is likely to be exac- change will push 100 million more people to factory disruption, increased demand, erbated by climate change, as the work- into poverty by 2030 due to factors like or transportation disruptions—and subse- force is also affected by the forces driving property loss, increased health burdens, quent resource shortages. migration. Quality of care is poor in many 24 and decreased crop yields. Poorer popu- Infrastructure damage limits a settings, with high rates of misdiagnosis lations are particularly susceptible to the facility’s ability to deliver essential, high and inappropriate treatments, which is 31 threats posed by climate change, creating quality services. People might be unable probably due to inadequate training. As a cycle in which climate change exacer- to access care due to transportation outlined in box 3, the workforce might be bates existing social and political issues by difficulties, caused by road damage or further impaired through cognitive effects both creating poverty and trapping people the unavailability of emergency medical of climate change and knowledge deficits, within it. services. Systems might face barriers to causing substantial problems in areas that Worsening poverty will contribute to maintaining public health and preventive already lack high quality training. higher burdens of disease, placing more strategies, such as the surveillance of The healthcare workforce is one of the stress on healthcare systems, and will put emerging threats. These many obstacles most important factors in UHC—both greater strains on government budgets for are likely to place additional cost burdens the availability of providers and the countries seeking to provide affordable, on health systems, which will trickle quality of care they provide. Shortages accessible care. down to the individual or insurance caused by geographic redistribution of

Box 1: Effects of climate change on disease burden Non-communicable diseases • A temperature rise of 1°C is linked to a 3.4% rise in cardiovascular mortality, a 3.6% rise in respiratory mortality, and a 1.4% rise in cerebrovascular mortality13 • High temperatures are linked to a 6% increase in hospital admissions for coronary artery disease14; cardiovascular events are also associated with exposure to air pollution—such as byproducts from the burning of fossil fuels (for example, fine particulate matter (PM2.5)) and ozone— which is amplified by temperature changes15 • Higher temperatures, increased intensity of wildfires, more severe and longer pollen seasons, and ground level air pollution like ozone and PM2.5 increase the burden of respiratory disease6 15 • Early data found that diabetes incidence increases by 0.314 per 1000 people for every 1°C rise in temperature,9 although more research is needed • Extreme weather, forced displacement, and violence can precipitate mental health concerns; extreme heat can exacerbate existing conditions11 • Chronic kidney disease of unknown origin has been linked to increasing heat stress in many regions, especially in agricultural communities16 Infectious diseases • Vectorial capacity for the transmission of malaria has increased by over 20% in higher elevations in Africa since 1950.3 WHO predicts major future rises in mortality due to climate change related increases in malaria in central and eastern regions of sub-Saharan Africa4 • Since the 1950s, vectorial capacity for the transmission of dengue has increased by 7.8-9.6%3 • Warmer ambient temperatures have been associated with foodborne illnesses, like salmonella17 • A 1°C rise in temperature may lead to a 0.8-2.1% increase in hand, foot, and mouth disease18 Occupational injuries • At higher temperatures, risk of work related injuries and illnesses increases among both indoor and outdoor workers19 • Outdoor workers in particular face increased risk of heat related illness as heatwaves become more frequent and last longer20

8 doi: 10.1136/bmj.l5302 | BMJ 2019;366:l5302 | the bmj UNIVERSAL HEALTH COVERAGE

Box 2: Effects of climate change on healthcare systems Low health risk country: hurricane Maria in the United States Hurricane Maria struck the US in 2017, causing major damage to Puerto Rico. Devastating rainfall and landslides, both attributed to climate change, battered the island.25 Combined with ineffective governance and disaster relief from the US government, the storm took a major toll on the island. The initial death count was 64, but subsequent household survey estimates place the loss of life at nearly 5700.26 The average household went 84 days without electricity, 68 days without water, and 41 days without cellular phone coverage.26 Nearly a third of households reported a disruption to health services, including 14% who were unable to access medications, 10% who were unable to use medical device equipment that required power, 9% who reported closed medical facilities, and 6% who found a lack of doctors. Nearly 9% of the most remote households could not reach emergency medical services by phone.26 Healthcare disruption was not limited to Puerto Rico. Approximately 44% of US intravenous fluid was produced on this island, and damage to the factory led to shortages that lasted months at hospitals throughout the US and other countries.27 The cascading effects of this one storm show how vulnerable supply chains are and the importance of fortifying climate resilience in healthcare systems. High health risk country: Bangladesh and flooding of the Brahmaputra River Bangladesh is a low lying, densely populated country with enormous vulnerability to rising sea levels and climate driven extreme weather, including flooding, cyclones, and drought.28 In 2017, climate change (along with other factors, such as deforestation and higher population density) contributed to the flooding of the Brahmaputra River, which destroyed nearly 500 community health clinics. This is catastrophic in a country with approximately five medical doctors and eight hospital beds per 10 000 people.29 The destruction of already limited healthcare infrastructure had a drastic effect on access to care. Solutions are difficult, as the climate driven weather is constantly creating and destroying floating islands, which makes establishing permanent clinics or hospitals impossible. Forced to develop innovative solutions, some communities have launched “floating hospitals.”28 These consist of boats that bring basic medical services to people who live in the riverine islands. A boat can see about 60 000 patients a year over hundreds of miles. Countries already at extreme risk, like Bangladesh, represent early examples of the enormous climate change related healthcare challenges we face and the need for creative solutions. providers might hinder access to care, will probably be financially beneficial in UHC and climate action agendas. These two whereas inadequate training might lead the long run. Some countries might ben- communities need open dialogue with each to misdiagnosis, ineffective disease efit more than others, but all can make other—with cross sectorial representation— surveillance, and, ultimately, harm to immense gains from taking an integrated and to push jointly for bold and innovative patients. systems approach. solutions.

Vulnerable countries Integrated solutions Novel financial frameworks The countries that are most vulnerable to Countries have already taken important Financial limitations are often seen as a climate change35 are often those that face steps towards tackling climate change major barrier to climate action, but this the greatest barriers to achieving UHC (fig through the Paris Agreement, which was field actually provides a major opportu- 2). This is not surprising—both are related called “the strongest health agreement nity for economic benefit. We need forward to the country’s economic strength and of this century” by WHO and outlines the thinking financial solutions to stimulate availability of resources. Because of the benefits of climate mitigation for health action in conjunction with political will. By unique challenges they face, these regions and development.6 Global leaders can and 2030, climate action is estimated to create have enormous opportunity to take a more must incorporate climate related threats $23 trillion (£19 trillion; €21 trillion) in integrated approach to their agendas. This into their considerations related to UHC investment opportunity in just 21 emerg- would not be easy; given the financial (box 4). ing market economies21—and this may restraints many of these countries already translate to healthcare. We also need to face related to healthcare, they might see Improved understanding and integrated include health and healthcare climate bur- tackling climate change as impossible. agendas dens in discussions around the economics Implementing and optimising UHC, how- New research to facilitate data driven solu- of climate action. As of February 2019, 46 ever, is a key strategy to minimise the tions would be helpful, but we already have national and 28 subnational jurisdictions health burdens of climate change and sufficient understanding to integrate the have implemented or planned carbon pric-

Box 3: Effects of climate change on the health workforce Climate sensitive health and travel concerns Healthcare workers might provide lower quality of care if they have impaired cognitive function due to climate change (extreme heat,32 nutritional deficiencies, and infectious diseases). Heat is of particular concern where air conditioning is permanently or frequently unavailable or water for cooling is scarce. Healthcare workers might also have difficulty reporting to work during extreme weather situations, such as when flooding disrupts roadways. Climate change and health knowledge deficits Climate change alters existing disease burdens through various routes. Although a growing proportion of the healthcare workforce recognises that climate change negatively affects health, there are gaps in understanding around the details.33 34 Only 19% of disease control workers in China understood that poor people were at greater risk of climate change related health problems. In addition, only one-third had a good understanding of how climate change affects the transmission of infectious diseases.34 the bmj | BMJ 2019;366:l5302 | doi: 10.1136/bmj.l5302 9 UNIVERSAL HEALTH COVERAGE

Climate vulnerability efforts, so appropriate education on local climate health is critical to improve their adaptive capacity.

Health system climate resiliency As climate change exacerbates existing threats and exposes new vulnerabilities, health systems must introduce forward thinking, data driven resiliency measures that are based on the unprecedented chal- lenges of the future. This mandates global assessments of climate hazards and health system vulnerabilities, which can then be tailored to unique local environments. The results of these assessments can then be implemented into strategic capital invest- ment priorities and appropriate health workforce management. Meanwhile, crea- tive workforce models and integration of UHC index technology will further bolster health sys- tem capacities.

Conclusions We are at an important point in time where action—or inaction—on the inter- secting issues of climate change and UHC will drive the health of nations for decades to come. Estimates show that we have about a decade to decrease green- house gas emissions to avoid the most catastrophic health outcomes.40 Thus, Very low the opportunities to transform health are Low enormous, and the time to act is now. As Moderate global decision makers aim to improve High Very High the health and quality of life for all peo- Not applicable ple, they must not overlook the effects that climate change will have on disease Fig 2 | A: Global map of climate change vulnerability. This map uses climate change burden and healthcare infrastructure. 35 vulnerability data from the Notre Dame Global Adaptation Initiative (ND-GAIN) Country Index, Only through bold, innovative, and cross which is based on indicators of adaptive capacity, sensitivity, and exposure and includes disciplinary action can we tackle these health, food, ecosystems, habitat, water, and infrastructure. Countries are categorised by unprecedented complex challenges and quintile. B: Global map of universal health coverage. This map uses the universal health ensure a healthier world for future gen- coverage index of essential service coverage data from the World Health Organization,36 which is based on indicators for reproductive, maternal/newborn/child health, infectious diseases, erations. non-communicable diseases, service capacity, and access. Countries are categorised by We thank Kathryn Horneffer and Robert J Redelmeier for their assistance. quintile. Contributors and sources: RNS is an emergency ing, with novel opportunities to integrate Adaptation to climate change medicine physician and researcher focused on the 21 health and healthcare system effects of climate with UHC (box 4). UHC is itself a fundamental adaptation change. AKJ is an internal medicine physician and intervention as it mitigates the negative researcher focused on quality and cost effectiveness Climate change mitigation health burdens of climate change. Mean- of healthcare, both globally and in the United A rapid transition to renewable energy, while, we need investment in research to States. Both authors contributed to the conception, preparation, review, and approval of this manuscript. which is both feasible and cost effective, understand the health risks of climate The sources of information used to create this would have direct health benefits now change in local populations. There then manuscript include the peer-reviewed literature and and would minimise health burdens in needs to be political and fiscal support to governmental or agency reports. AKJ is the guarantor. the future. But we need political will to translate this research into interventions Patient involvement: No patients were involved. take the urgent and bold steps necessary. and infrastructure that protect the most Competing interests: We have read and understood Mitigation must also specifically occur in vulnerable. Another essential component BMJ policy on declaration of interests and have no the health sector, which contributes to a of achieving UHC is the development of a competing interests to declare. disproportionate amount of carbon emis- dynamic health workforce that can respond This article is part of a series commissioned by The sions.39 Health professionals can play an BMJ based on an idea from the Harvard Global Health to the changing needs of a region. Health- Institute. The BMJ retained full editorial control over important role in advocating for policies care workers will be on the front lines of external peer review, editing, and publication. Harvard that will incentivise this transition. disaster response and disease surveillance Global Health Institute paid the open access fees.

10 doi: 10.1136/bmj.l5302 | BMJ 2019;366:l5302 | the bmj UNIVERSAL HEALTH COVERAGE

Box 4: Sample solutions that tackle both the universal health coverage (SDG 3.8) and climate action (SDG 13) Improved understanding and integrated agendas • Understand how individuals use healthcare for climate sensitive conditions and determine how different UHC models will affect this • Gather experts across sectors (as in the One Health approach) to develop an agenda for tackling these issues together • Develop joint metrics tracking both SDGs (3 and 13) Novel financial frameworks • Use carbon pricing that includes climate driven health and healthcare system costs • Finance UHC from the elimination of fossil fuel subsidies37 and carbon pricing Climate change mitigation • Frame transition to renewable energy around the anticipated health and health equity benefits • Ensure that transition to renewable energy is urgent and extensive in healthcare facilities, fuelled by advocacy from healthcare professionals and political leaders • Broad divestment from fossil fuel companies to numerous sectors, especially healthcare Adaptation to climate change • Data driven approach to identifying those most vulnerable to heat exposure in a city, how they access care, and how the public health infrastructure can best protect them through adaptation interventions • Translate data into effective surveillance systems and efficient sharing of emerging health concerns across borders • Train medical professionals in skills that transcend current specialty boundaries, such as disaster preparedness training for hospitalists, and knowledge of emerging climate sensitive threats, such as new geographic distributions of infectious diseases Health system resilience • Map out climate hazards, such as flooding and other extreme weather implications, for local regions using different future climate models (eg, moderate to severe) • Redesign facilities (eg, protection from flooding), relocate generators (eg, roof placement), and engage with the local health community (eg, coordination between local hospitals) • Create incentives for strategic geographic placement of the health workforce and use health technologies like predictive staffing models and telemedicine

Renee N Salas, affiliated faculty 6 World Health Organization (WHO). COP24 special health. Lancet 2015;386:1861-914. doi:10.1016/ Ashish K Jha, faculty director report: health and climate change. World Health S0140-6736(15)60854-6 Organization, 2018. 16 Sorensen C, Garcia-Trabanino R. A new era of climate Harvard Global Health Institute, 42 Church St, 7 United States Environmental Protection Agency medicine—addressing heat-triggered renal disease. Cambridge, MA 02138, USA (EPA). Global greenhouse gas emissions data. N Engl J Med 2019;381:693-6. doi:10.1056/ Correspondence to: A K Jha https://www.epa.gov/ghgemissions/global- NEJMp1907859 [email protected] greenhouse-gas-emissions-data 17 Milazzo A, Giles LC, Zhang Y, Koehler AP, Hiller JE, 8 Bloom G, Katsuma Y, Rao KD, Makimoto S, Yin Bi P. Factors influencing knowledge, food safety JDC, Leung GM. Next steps towards universal practices and food preferences during warm weather health coverage call for global leadership. of Salmonella and Campylobacter cases in South BMJ 2019;365:l2107. doi:10.1136/bmj.l2107 Australia. Foodborne Pathog Dis 2017;14:125-31. This is an Open Access article distributed in 9 Blauw LL, Aziz NA, Tannemaat MR, et al. Diabetes doi:10.1089/fpd.2016.2201 accordance with the Creative Commons Attribution incidence and glucose intolerance prevalence 18 Wei J, Hansen A, Liu Q, Sun Y, Weinstein P, Bi P. Non Commercial (CC BY-NC 4.0) license, which increase with higher outdoor temperature. BMJ Open The effect of meteorological variables on the permits others to distribute, remix, adapt, build Diabetes Res Care 2017;5:e000317. doi:10.1136/ transmission of hand, foot and mouth disease in upon this work non-commercially, and license bmjdrc-2016-000317 four major cities of Shanxi province, China: a time their derivative works on different terms, provided 10 World Health Organization (WHO). World health series data analysis (2009-2013). PLoS Negl Trop the original work is properly cited and the use is statistics 2018: monitoring health for the Dis 2015;9:e0003572. doi:10.1371/journal. non-commercial. See: http://creativecommons.org/ sustainable development goals (SDGs). World Health pntd.0003572 licenses/by-nc/4.0/. Organization, 2018. 19 Varghese BM, Barnett AG, Hansen AL, et al. 11 Berry HL, Bowen K, Kjellstrom T. Climate change The effects of ambient temperatures on the and mental health: a causal pathways framework. risk of work-related injuries and illnesses: Int J Public Health 2010;55:123-32. doi:10.1007/ Evidence from Adelaide, Australia 2003-2013. s00038-009-0112-0 Environ Res 2019;170:101-9. doi:10.1016/j. 1 Hickel J. The contradiction of the sustainable 12 Wondmagegn BY, Xiang J, Williams S, Pisaniello D, Bi envres.2018.12.024 development goals: Growth versus ecology on P. What do we know about the healthcare costs of 20 Varghese BM, Hansen A, Bi P, Pisaniello D. Are a finite planet [in press]. Sustain Dev 2019. extreme heat exposure? A comprehensive literature workers at risk of occupational injuries due doi:10.1002/sd.1947 review. Sci Total Environ 2019;657:608-18. to heat exposure? A comprehensive literature 2 World Health Organization (WHO), The World Bank. doi:10.1016/j.scitotenv.2018.11.479 review. Saf Sci 2018;110:380-92. doi:10.1016/j. Tracking universal health coverage: 2017 global 13 Bunker A, Wildenhain J, Vandenbergh A, et al. ssci.2018.04.027 monitoring report. World Health Organization, 2017. Effects of air temperature on climate- 21 The World Bank. Climate change. https://www. 3 Watts N, Amann M, Arnell N, et al. The 2018 report of sensitive mortality and morbidity outcomes worldbank.org/en/topic/climatechange/overview the Lancet Countdown on health and climate change: in the elderly: a systematic review and 22 Bowles DC, Butler CD, Morisetti N. Climate change, shaping the health of nations for centuries to come. meta-analysis of epidemiological evidence. conflict and health. J R Soc Med 2015;108:390-5. Lancet 2018;392:2479-514. doi:10.1016/S0140- EBioMedicine 2016;6:258-68. doi:10.1016/j. doi:10.1177/0141076815603234 6736(18)32594-7 ebiom.2016.02.034 23 Stapleton S, Nadin R, Watson C, Kellett J. Climate 4 World Health Organization (WHO). Climate and 14 Bai L, Li Q, Wang J, et al. Increased coronary heart change, migration and displacement: the need for health country profiles: a global overview. World disease and stroke hospitalisations from ambient a risk-informed and coherent approach. Overseas Health Organization, 2015. temperatures in Ontario. Heart 2018;104:673-9. Development Institute, 2017. 5 Haines A, Ebi K. The imperative for climate action doi:10.1136/heartjnl-2017-311821 24 Hallegatte S, Bangalore M, Bonzanigo L, et al. Shock to protect health. N Engl J Med 2019;380:263-73. 15 Watts N, Adger WN, Agnolucci P, et al. Health and Waves: Managing the Impacts of Climate Change on doi:10.1056/NEJMra1807873 climate change: policy responses to protect public Poverty. World Bank Group, 2016. the bmj | BMJ 2019;366:l5302 | doi: 10.1136/bmj.l5302 11 UNIVERSAL HEALTH COVERAGE

25 Keellings D, Ayala JJH. Extreme rainfall associated Health (Abingdon) 2006;19:385-7. Province, China. PLoS One 2014;9:e109476. with hurricane Maria over Puerto Rico and doi:10.1080/13576280600937911 doi:10.1371/journal.pone.0109476 its connections to climate variability and 31 Das J, Woskie L, Rajbhandari R, Abbasi K, Jha 35 University of Notre Dame. ND-GAIN: Notre Dame change. Geophys Res Lett 2019;46:2964-73. A. Rethinking assumptions about delivery of Global Adaptation Initiative. https://gain.nd.edu/ doi:10.1029/2019GL082077 healthcare: implications for universal health 36 World Health Organization (WHO). Universal health 26 Kishore N, Marqués D, Mahmud A, et al. coverage. BMJ 2018;361:k1716. doi:10.1136/ coverage index of essential service coverage (%). Mortality in Puerto Rico after Hurricane Maria. bmj.k1716 http://apps.who.int/gho/data/node.imr.UHC_INDEX_ N Engl J Med 2018;379:162-70. doi:10.1056/ 32 Cedeño Laurent JG, Williams A, Oulhote Y, Zanobetti REPORTED?lang=en NEJMsa1803972 A, Allen JG, Spengler JD. Reduced cognitive function 37 Gupta V, Dhillon R, Yates R. Financing universal health 27 Salas RN, Knappenberger P, Hess J. Lancet during a heat wave among residents of non-air- coverage by cutting fossil fuel subsidies. Lancet countdown on health and climate change brief for conditioned buildings: an observational study Glob Health 2015;3:e306-7. doi:10.1016/S2214- the United States of America. Lancet Countdown, of young adults in the summer of 2016. PLoS 109X(15)00007-8 2018. Med 2018;15:e1002605. doi:10.1371/journal. 39 Pichler P-P, Jaccard IS, Weisz U, Weisz H. International 28 Sampath N. Floating hospitals treat those pmed.1002605 comparison of health care carbon footprints. Environ impacted by rising seas. Natl Geogr 2017. https:// 33 Wei J, Hansen A, Zhang Y, et al. Perception, attitude Res Lett 2019;14:064004. doi:10.1088/1748- www.nationalgeographic.com/news/2017/03/ and behavior in relation to climate change: a survey 9326/ab19e1 . floating-hospitals-bangladesh-climate-change- among CDC health professionals in Shanxi province, 40 Intergovernmental Panel on Climate Change (IPCC). refugees/ China. Environ Res 2014;134:301-8. doi:10.1016/j. Special report on global warming of 1.5°C (SR15). 29 Global Health Observatory. Countries: Bangladesh. envres.2014.08.006 2018. https://www.ipcc.ch/sr15/ https://www.who.int/countries/bgd/en/ 34 Wei J, Hansen A, Zhang Y, et al. The impact of 30 Guilbert JJ. The World Health Report climate change on infectious disease transmission: Cite this as: BMJ 2019;366:l5302 2006: working together for health. Educ perceptions of CDC health professionals in Shanxi http://dx.doi.org/10.1136/bmj.l5302

12 doi: 10.1136/bmj.l5302 | BMJ 2019;366:l5302 | the bmj UNIVERSAL HEALTH COVERAGE

Achieving universal health coverage for mental disorders Vikram Patel and Shekhar Saxena discuss strategies for integrating mental disorders, and other chronic conditions, into primary healthcare to achieve universal health coverage

ental disorders are the lead- tiveness of the coverage) are even larger. disorders (or any chronic condition).8 ing contributor to the global Across all income categories, countries Historically, primary healthcare was for burden of years lived with invest tiny fractions of their healthcare acute or episodic medical events—from disability.1 The burden and budgets on mental health, disproportion- cuts to colds, extending to childbirth cost of untreated mental dis- ately less than the burden of mental disor- and the management of acute infections Morders is immense for individuals, families, ders.5 This results in an inadequate number such as malaria and diarrhoea. Anything communities, and ultimately the world.2 of mental health professionals per capita, a more complex (involving long term care A key strategy to tackling this burden is massive shortage of community based men- or requiring a person centred approach universal coverage of cost effective inter- tal healthcare, and the persistence of badly to care, beyond a reductionist biomedical ventions for mental disorders,3 one target run large mental hospitals. diagnosis) was either ignored or passed of the sustainable development goals. We These barriers to supply are compounded to secondary care. Attempts to integrate analyse key questions related to this goal: by barriers to demand, related to stigma mental health have failed because they why have countries failed to achieve univer- and the discrepancies between biomedical have not tackled these fundamental sal health coverage for mental disorders? framing of mental health problems and the barriers; instead, they have tried to How can mental health interventions be conceptualisation of emotional distress replicate secondary care in primary care— integrated in primary healthcare, the foun- in the community. Barriers to demand for example, by posting psychiatrists in 9 10 dational platform of delivery of universal are one of the reasons for the large gaps primary healthcare centres, a strategy health coverage? What are the lessons for in coverage observed in well resourced that is neither scalable nor necessary. integrating other chronic conditions into contexts, where universal supply of mental primary healthcare? healthcare interventions has been largely How can mental healthcare be integrated into attained—such as in the UK with its diverse primary care? Why have countries failed to achieve universal mental healthcare programmes including A key element of the field of global men- mental health coverage? community based mental healthcare teams tal health is the design and evaluation of and the Improving Access to Psychological In the context of mental healthcare, all innovative strategies for integrating cost Therapies programme. Further, national effective pharmacological and psycho- countries are “developing” to some extent. averages hide enormous inequities social interventions in primary health- Even in high income countries, the cover- within countries, both geographic and care.1 11 The evidence from this work, from age gaps for common conditions like mood societal—indigenous, minority, rural, and a range of contexts including high income and anxiety disorders often exceed 50%; socially and economically disadvantaged countries, is showing the way to integra- in low income countries, the gap exceeds communities have much poorer access tion.1 A theme across this evidence is the 90%.4 Quality gaps (a measure of the effec- to quality care. A particularly egregious placement of non-specialised providers example is people with severe mental (including peers, community health work- KEY MESSAGES disorders experiencing a loss of up to ers, and nurses) in primary healthcare and All countries have failed to achieve half their life expectancy relative to community settings to perform diverse roles • 6 12 universal health coverage for mental the general population, being more such as coordinating collaborative care ; disorders, owing to barriers related to likely to experience homelessness and educating and mobilising the community 13 budget and stigma marginalisation, and being denied the to increase demand for care ; supporting The architecture of primary health- basic rights to freedom and dignity through families and patients to tackle proximal • 7 care in most countries is simply not fit incarceration in hospitals or prisons. social determinants of mental health; and for the integration of mental disorders Since the Alma Ata declaration in 1978, delivering empirically supported psycho- 14 (or any chronic condition) the means of improving access to mental logical and social interventions. healthcare has been to integrate it with The growing recognition that binary • Integration can be achieved using an approach that is person centred, col- primary healthcare. But after four decades models of diagnosis of mental disorders do laborative, compassionate, engaged of trying, we know that achieving such not capture the dimensional distribution with the community, and that integration at scale will require nothing of symptoms, distress, and disability of includes long term planning short of a wholescale re-engineering of mental health problems in the population the healthcare system. At the heart of the has important implications for treatment These principles could be applied • challenge is the architecture of primary planning. A “one size fits all” approach across the full range of chronic con- healthcare in most countries, which is does not work. Instead, we need a ditions simply not fit for the integration of mental staged approach whereby interventions the bmj | BMJ 2019;366:l4516 | doi: 10.1136/bmj.l4516 13 UNIVERSAL HEALTH COVERAGE

are delivered based on both symptom to seek care, less likely to use digital to promote mental health. Examples of severity and the effect of these symptoms technology, and more likely to experience innovative delivery strategies include using on distress and disability.15 This aligns discrimination, isolation, and premature non-specialist providers, including peers, with the notion that most care targeting mortality. Second, coverage must be equity to deliver psychosocial interventions,14 relatively mild, early, or transient stages of sensitive, recognising that subgroups in the using digital platforms to support guided mental distress can be delivered through population that experience higher levels self care and training and supervision of self care and by non-specialist providers of deprivation or exclusion, such as poor providers,16 and using collaborative care with appropriate skills. This facilitates people, refugees or ethnic, religious, or with case managers to manage multiple quicker recovery for those people while sexual minorities, bear a disproportionate morbidities.12 18 Engaging civil society to simultaneously identifying those who burden of mental disorders. Third, increase the demand for care, to tackle might need more intensive interventions integration must emphasise quality of care stigma and discrimination, and to design, for referral to specialised providers. This for both the mental disorder (for example, deliver, and hold services accountable has approach, which reduces the emphasis on to abolish coercive, harmful, and abusive helped reduce barriers to demand while also biomedical diagnoses, is also more likely to practices) and co-existing physical health empowering people with lived experience. be acceptable to the wider population and conditions, which are major contributors to Despite robust evidence from pilot to be less stigmatising. premature mortality. studies and trials, little progress has been The exponential growth of digital health made in scaling up these strategies in most innovations—spanning guided self care, What are the lessons for integrating chronic countries.19 Major barriers that remain training and supervision of frontline conditions? include financing of non-specialist providers workers, remote consultations by specialist Mental healthcare has led the develop- to deliver psychosocial interventions; providers, and remote monitoring of mental ment of care strategies for health condi- implementing scalable approaches to health—offers a transformative opportunity tions characterised by a chronic, episodic, training, supervision, support, and to bypass historical structural barriers to or relapsing course. At the heart of these quality assurance; and institutionalising enabling task sharing and collaborative innovations is the transition of delivery collaborative and coordinated care. The care.16 Several examples of innovative of long term care from institutions to the goal of improving the recognition of mental programmes and projects can be found community, with the goal of decreasing dis- disorders and delivery of mental health in the Mental Health Innovation Network abilities, optimising quality of life, slowing interventions in primary care remains a (www.mhinnovation.net) and the Lancet disease progression, and minimising the distant one for most of the world. Commission on Global Mental Health and risk of relapse. Interventions have focused This focus on implementation science Sustainable Development.17 on “recovery” by going beyond the specific is at the heart of the work of the PRIME We have identified five key elements symptoms of the disorder to tackle impair- consortium20 (sponsored by the UK needed to integrate mental health in ments in daily life and experiences priori- Department for International Development) primary care, which we refer to as the tised by the patient—the hallmark of person and the National Institute of Mental “5C approach” (box 1). Although these centred care. Health’s research partnerships for scaling elements were derived from innovations Mental health programmes have up mental health interventions in low and seeking to attain universal coverage of championed the integration of middle income countries (https://www. mental healthcare, they can be applied to pharmacological with psychological and nimh.nih.gov/about/organization/cgmhr/ the full range of chronic conditions. social interventions, referred to as the scaleuphubs/index.shtml). The goal is Three key points need to be emphasised. biopsychosocial approach to care; the to show how routine healthcare systems First, integration must cover the full range engagement of family members (where can fully integrate the strategies that of mental disorders, in particular ensuring culturally appropriate and agreed with the have proved effective for the management that people with severely disabling patient) to support recovery and tackling of mental disorders. A key strategy is conditions—such as schizophrenia, alcohol the needs of caregivers; and intersectoral integration of care with other chronic and drug dependence, and dementia— interventions to promote the inclusion conditions. Mental disorders (including are not left behind as they are less likely of people with mental disabilities and substance use disorders), for example, worsen the outcomes of cardiovascular and metabolic disorders, which are major Box 1: The “5C” approach to integrating mental health in universal health coverage contributors to the premature mortality of • Person centred—focusing on what matters to the patient rather than what is the matter with people with severe mental disorders.18 The the patient,21 which translates into attending to functional needs, multiple morbidities, synergies between non-communicable and social suffering and to the empowerment of the person to harness their personal and diseases and mental health problems are community resources to enable recovery with dignity recognised in inclusion of mental health in • Continuing or long term planning—recognising that “cures” are rare and the goal of care is to the scope of the World Health Organization’s optimise the quality of life and health independent high level commission and the • Community platform of delivery—engagement with families and the broader community to United Nations’ high level meeting on non- tackle stigma, adherence, and other barriers to the uptake of effective care1 communicable diseases. • Collaborative care—with seamless coordination by community health workers or case managers of primary care and specialist providers to ensure high coverage of quality care and Conclusions early “stepping up” of the intensity of care when needed12 As the global health community reflects • Compassionate stance—instils hope, a key ingredient for patient engagement, motivates on the role of primary healthcare in this health promoting behaviours, and harnesses the placebo effect (which has a robust 40th anniversary of the historic Alma Ata neuroscientific basis)22 declaration, we reaffirm the view that pri-

14 doi: 10.1136/bmj.l4516 | BMJ 2019;366:l4516 | the bmj UNIVERSAL HEALTH COVERAGE

mary healthcare must be the foundation Non Commercial (CC BY-NC 4.0) license, which 11 Collins PY, Saxena S. Action on mental health of the architecture of universal health cov- permits others to distribute, remix, adapt, build needs global cooperation. Nature 2016;532:25-7. upon this work non-commercially, and license doi:10.1038/532025a erage to realise the goals of reducing the their derivative works on different terms, provided 12 Patel V, Belkin GS, Chockalingam A, et al. unmet need for mental healthcare glob- the original work is properly cited and the use is Integrating mental health services into priority ally. Achieving this, however, will require non-commercial. See: http://creativecommons.org/ health care platforms: addressing a grand licenses/by-nc/4.0/. challenge in global mental health. PLoS fundamental re-engineering of the way Med 2013;10:e1001448. doi:10.1371/journal. that primary healthcare is conceptualised, pmed.1001448 organised, and delivered, and this, in turn, 13 Shidhaye R, Murhar V, Gangale S, et al. The effect of VISHRAM, a grass-roots community-based will need the full engagement and support 1 Patel V, Chisholm D, Parikh R, et al, DCP MNS mental health programme, on the treatment of all actors in universal healthcare, not Author Group. Addressing the burden of mental, gap for depression in rural communities least people who are affected by mental neurological, and substance use disorders: key in India: a population-based study. Lancet disorders. The rising burden of mental dis- messages from Disease Control Priorities, 3rd Psychiatry 2017;4:128-35. doi:10.1016/S2215- edition. Lancet 2016;387:1672-85. doi:10.1016/ 0366(16)30424-2 orders, in all countries, requires immediate S0140-6736(15)00390-6 14 Singla DR, Kohrt BA, Murray LK, Anand A, and dramatic actions, informed by the rich 2 Chisholm D, Sweeny K, Sheehan P, et al. Scaling- Chorpita BF, Patel V. Psychological treatments body of evidence on delivery innovations up treatment of depression and anxiety: a for the world: lessons from low- and global return on investment analysis. Lancet middle-income countries. Annu Rev Clin from diverse contexts. Failure to do so will Psychiatry 2016;3:415-24. doi:10.1016/S2215- Psychol 2017;13:149-81. doi:10.1146/annurev- mean failure to achieve universal health 0366(16)30024-4 clinpsy-032816-045217 coverage—universal refers not just to cov- 3 Chisholm D, Naci H, Hyder AA, Tran NT, Peden M. 15 Patel V. Talking sensibly about depression. PLoS Cost effectiveness of strategies to combat road traffic erage of the population but also coverage Med 2017;14:e1002257. doi:10.1371/journal. injuries in sub-Saharan Africa and South East Asia: pmed.1002257 of the full range of its health needs, and mathematical modelling study. BMJ 2012;344:e612. 16 Naslund JA, Aschbrenner KA, Araya R, et al. Digital health should be considered comprehen- doi:10.1136/bmj.e612 technology for treating and preventing mental sively, to include mental health and social 4 Thornicroft G, Chatterji S, Evans-Lacko S, disorders in low-income and middle-income et al. Undertreatment of people with major countries: a narrative review of the literature. Lancet wellbeing alongside physical health. depressive disorder in 21 countries. Br J Psychiatry 2017;4:486-500. doi:10.1016/S2215- Psychiatry 2017;210:119-24. doi:10.1192/bjp. 0366(17)30096-2 Contributors and sources: Both authors came up bp.116.188078 17 Patel V, Saxena S, Lund C, et al. The Lancet with the concept of the article. VP led the drafting and 5 World Health Organization. Mental Health ATLAS Commission on global mental health and sustainable SS reviewed, edited, and contributed to the draft. Both 2014. WHO, 2015. development. Lancet 2018;392:1553-98. authors have approved the revision. 6 Walker ER, McGee RE, Druss BG. Mortality in mental doi:10.1016/S0140-6736(18)31612-X Competing interests: Neither author has any disorders and global disease burden implications: 18 Patel V, Chatterji S. Integrating mental health conflicts of interest to disclose. VP is supported a systematic review and meta-analysis. JAMA in care for noncommunicable diseases: an by National Institute of Mental Health and the UK Psychiatry 2015;72:334-41. doi:10.1001/ imperative for person-centered care.Health Aff Department for International Development grants. jamapsychiatry.2014.2502 (Millwood) 2015;34:1498-505. doi:10.1377/ Provenance and peer review: Commissioned; 7 Kleinman A. Global mental health: a failure of hlthaff.2015.0791 externally peer reviewed. humanity. Lancet 2009;374:603-4. doi:10.1016/ 19 Eaton J, McCay L, Semrau M, et al. Scale up of S0140-6736(09)61510-5 services for mental health in low-income and middle- This article is part of a series commissioned by The 8 Beaglehole R, Epping-Jordan J, Patel V, et al. income countries. Lancet 2011;378:1592-603. BMJ based on an idea from the Harvard Global Health Improving the prevention and management of doi:10.1016/S0140-6736(11)60891-X Institute. The BMJ retained full editorial control over chronic disease in low-income and middle-income 20 Lund C, Tomlinson M, De Silva M, et al. PRIME: a external peer review, editing, and publication. Harvard countries: a priority for primary health care. programme to reduce the treatment gap for mental Global Health Institute paid the open access fees. Lancet 2008;372:940-9. doi:10.1016/S0140- disorders in five low- and middle-income countries. Vikram Patel, Pershing Square professor of global 6736(08)61404-X PLoS Med 2012;9:e1001359. doi:10.1371/journal. health and Wellcome Trust principal research fellow 9 Gilbody S, Whitty P, Grimshaw J, Thomas R. pmed.1001359 Shekhar Saxena, professor of practice Educational and organizational interventions to 21 Kebede S. Ask patients “What matters to you?” rather improve the management of depression in primary Harvard TH Chan School of Public Health, USA than “What’s the matter?”BMJ 2016;354:i4045. care: a systematic review. JAMA 2003;289:3145-51. doi:10.1136/bmj.i4045 Correspondence to: V Patel doi:10.1001/jama.289.23.3145 22 Geuter S, Koban L, Wager TD. The cognitive [email protected] 10 van Ginneken N, Maheedhariah MS, Ghani neuroscience of placebo effects: concepts, S, Ramakrishna J, Raja A, Patel V. Human predictions, and physiology. Annu Rev resources and models of mental healthcare Neurosci 2017;40:167-88. doi:10.1146/annurev- integration into primary and community care neuro-072116-031132 in India: Case studies of 72 programmes. PLoS This is an Open Access article distributed in One 2017;12:e0178954. doi:10.1371/journal. Cite this as: BMJ 2019;366:l4516 accordance with the Creative Commons Attribution pone.0178954 http://dx.doi.org/10.1136/bmj.l4516

the bmj | BMJ 2019;366:l4516 | doi: 10.1136/bmj.l4516 15 UNIVERSAL HEALTH COVERAGE

Motivating provision of high quality care: it is not all about the money Mylène Lagarde, Luis Huicho, and Irene Papanicolas discuss different strategies policy makers can use to motivate health providers in order to improve quality of care

he inclusion of universal health lines.10 11 Here, we discuss the evidence are reimbursed differently for inputs with coverage as a target in the sus- on different approaches that can be used similar health benefits they may choose tainable development goal for to increase provider motivation and ulti- the more profitable ones at the expense health has boosted the global mately improve quality of care. of efficiency.20 More generally, questions movement to improve access to remain about the cost effectiveness of healthcare services. To improve health, Financial incentives are not always worth the using financial incentives, especially when T 21 the services accessed must be high qual- investment budgets are tight. ity,1 yet there is mounting evidence that the Economists, managers, and policy makers The multiple ways in which financial quality of care delivered to populations in have long seen remuneration as an obvi- incentives can be designed, as well as the many low and middle income countries is ous lever to influence providers’ behaviour. complexity of the healthcare environment inadequate.2-5 Governments must consider In high income settings, the use of direct in which they are introduced,22 may help strategies that will not only improve acces- financial incentives to improve quality has explain why results have not always been sibility to care for their populations but also been ubiquitous with the aim of maintain- as expected.23 Incentive schemes can substantially improve quality. ing high quality standards while encour- differ in terms of the number and types of A priority in achieving universal health aging more efficient spending. In settings performance indicators targeted, the size of coverage is the recruitment, training, and where salaries are low and health workers the reward in relation to provider’s income, retention of healthcare workers. However, demotivated,7 similar pay-for-performance or the extent to which the performance there is widespread concern that health schemes have been used, often to achieve targets are completely under a provider’s systems are not getting the most out of their a dual objective: to increase remuneration control. All of these design choices workforce. Recent evidence shows that the and to provide incentives for improving matter and influence the effect of the quality of care provided by healthcare performance. incentives.24 The incentive will also be more workers is often lower than what they Despite the enthusiasm for these schemes, powerful if it directly targets individual are able to demonstrate in the context of the evidence of their impact on quality of providers (specialists) rather than small a test2 or under the watchful eyes of an care is lacklustre. Although they have been teams (primary care centres) or large observer.6 The existence of such “know-do” found to increase adherence to quality of organisations (hospitals). If individuals gaps shows that substandard care cannot care processes, their overall effects are mixed value losses more than gains, penalties for be fully explained by low competence or and, when positive, small.12-14 For example, failing to achieve targets will work better inadequate training. an experiment in Rwanda led to the than rewards for doing so.25 Low quality of care and medical errors improvement of some rewarded measures of Lastly, providers’ personal characteristics occur more often when providers are demo- process of care (eg, iron supplementation for may affect how they respond to incentives. tivated, which can be fuelled by inadequate children, urine analysis in antenatal care), For instance, a recent study found that working conditions such as shortages of but not of others (eg, malaria prophylaxis certain personality traits may predispose basic drugs and equipment or staff.7-9 Yet, and tetanus vaccination for pregnant some people to respond more to incentives although good working conditions are an women),15 16 and provider effort in antenatal than others.26 For people driven by factors important part of delivering good quality of consultations was only slightly higher than other than remuneration, financial care, they are not sufficient to ensure that in the absence of incentives.17 This is one incentives may backfire.27 This may be health professionals are motivated and of several examples of the mixed results particularly relevant to the healthcare adhere to recommended treatment guide- achieved by pay-for-performance schemes sector, as workers are likely to care about in terms of quality improvement in low not only their income but also their income settings, which are disappointing reputation, their patients, or their job. KEY MESSAGES considering the investment made.18 Reputation matters in some contexts • Policy makers need to look beyond The global evidence is similarly traditional financial incentives when mixed, with no evidence of significant Another way to incentivise quality improve- designing policies to improve care improvement in health outcomes after ment is to publish providers’ perfor- the introduction of pay-for-performance mance to the public or to their peers. This Health professionals are motivated by • schemes.13 Furthermore, caution is needed approach harnesses the power of another a range of factors, both extrinsic and to avoid unintended consequences that source of extrinsic motivation, concerns intrinsic can be costly or compromise quality. For for your reputation, which are expected to Incentives that focus on these other 25 28 • example, when income is directly linked push individuals to perform better. motivations or a combination may be to drug sales, more unnecessary drugs are Many countries have made measures of more effective 19 likely to be prescribed ; when providers hospital performance publicly available,

16 doi: 10.1136/bmj.l5210 | BMJ 2019;366:l5210 | the bmj UNIVERSAL HEALTH COVERAGE

typically reporting on waiting times or Harnessing the power of intrinsic motivation which attributes are associated with better patient experience, and sometimes on Intrinsic motivation is known to be a pow- patient care for specific types of healthcare measures of quality and safety, such as erful driver of provider behaviours.41-44 Two workers to inform such initiatives. mortality and complications. Most evidence different sources of motivation are identi- Selecting the “right” attributes is also on the impact of public reporting comes fied: the satisfaction derived from under- critical when recruiting people for specific from the US and shows mixed results.29 taking actions that benefit other people or jobs. As economic theory suggests that For hospitals, public reporting has been society (sometimes referred to as altruistic people sort into jobs matching their linked to small increases in adherence to or prosocial motivation) and the interest or preferences or motivation,53 54 employers processes of care30 but not to reductions enjoyment of a task itself. could attract the “right” people by in mortality.31 32 For individual health Evidence is emerging that intrinsically emphasising certain job characteristics. A professionals, it has been associated with motivated providers display desirable non-governmental organisation recently decreases in mortality,33 34 but there are behaviours or attitudes towards patients. tested this idea in Uganda, exploring concerns about how these reductions are Health education research in high income whether highlighting particular aspects achieved, and in particular the extent to countries shows the importance and long of the role of community health workers which public reporting encourages patient term benefits of selecting people with would make the job appeal to more selection.35 36 altruistic values, such as compassion or altruistic people.55 Information about job In settings where administrative empathy, into the medical profession.45 remuneration was manipulated to suggest to information on provider quality is lacking Research in sub-Saharan Africa has applicants that the main role was either more or less credible, public reporting of provider found that nurses who are more generous commercial (selling small products such as performance can rely on community towards patients are more likely to choose salt or soap to households) or more prosocial based monitoring of service delivery and jobs in rural and remote areas,46 and that (providing health advice). When lower engagement with providers.37 Robust more generous clinicians provide better remuneration was advertised, potential evidence on such initiatives is still limited, quality of care to patients.47 Yet, few applicants inferred the social aspect of the but in a pioneering experiment in Uganda, interventions have rigorously explored the job was more important; those who applied when non-governmental organisations extent to which intrinsic motivation can be showed higher levels of altruism and ended distributed reports on use and quality of shaped or harnessed to motivate quality up staying longer on the job and visited services to local communities, mortality in improvement. households more frequently.55 Similar children under 5 years old fell by a third.38 Policy interventions that could appeal success with framing job adverts has been Public reporting programmes could to this type of motivation fall into two found in other sectors,56 57 and it deserves therefore be a worthwhile strategy to categories, depending on whether intrinsic further exploration. promote quality, but their effectiveness motivation is seen to be malleable or not. If requires careful design and a favourable intrinsic motivation is an innate individual Nurturing intrinsic motivation through environment.22 The information reported trait, policy makers should try to select feedback has to be credible and salient. To achieve more people who display the right type of Few policies have sought to specifically this, it has to come from a trustworthy motivation. On the other hand, if intrinsic influence or nurture the motivational source; depending on the setting this motivation is a form of capital that can capital of providers. Examples include may or may not be the government.39 be depreciated or accrued,48 specific encouraging quality improvement through Equally, the information reported has to be interventions should be introduced to educational outreach programmes or audit noticeable and focused on a few indicators, nurture it. We consider recent examples of and feedback, either to promote the defini- to avoid being ignored.40 both approaches. tion of shared quality norms among groups Public reporting will improve quality of providers, or to highlight the benefits of only if providers are held accountable for Selecting people with intrinsic motivation good quality for patients. their performance in some way. Even in the Many countries have introduced pro- An example of such interventions absence of potential reputational or legal grammes to select people into medical or are physician collaboratives, which nursing training by introducing quotas have some positive effects on quality of consequences, publicly reporting poor 58 practice can have financial consequences based on people’s geographical origin in care. This approach uses confidential order to increase the rural retention of staff, performance feedback to individuals and in a competitive healthcare market. If 49 50 providers’ income is linked to the volume especially in low income settings, but then creates opportunities to share lessons of patients, low performers will be selection of medical students is still mostly in a multidisciplinary setting and agree automatically penalised as patients will based on academic attainment. There are on high quality standards and practices. “vote with their feet” and choose higher concerns that this approach is not sufficient The Swedish quality registers are often to ensure that people with softer skills are highlighted as an example that has had quality providers. But to achieve this, 51 59 patients must be able to both access and selected. A few countries have started to a sustained and large effect on quality. use the information reported. In many introduce selection procedures to identify Crucial to their success seems to be the settings, this may require additional medical workers with personality traits and fact that they are not part of an external values indicative of intrinsic and altruis- regulatory or performance management supporting initiatives to help patients 52 navigate a complex environment, or simply tic motivation. However, such initiatives process, but that they are driven by use and interpret the data. Conversely, are still in their infancy, and controversy physicians themselves, who have promoted remains about the types of non-academic a culture of constructive appraisal and if poor practice does not lead to negative 59 consequences for providers—whether attributes that should be included and the commitment to quality. financial, moral or legal—public reporting validity of the selection procedures. More Whereas physician collaboratives will be a toothless policy. research needs to be carried out to determine highlight the importance of professional the bmj | BMJ 2019;366:l5210 | doi: 10.1136/bmj.l5210 17 UNIVERSAL HEALTH COVERAGE

norms and expectations to foster intrinsic in mind that health professionals are Irene Papanicolas, associate professor of health economics1,3 motivation, two other examples show how heterogeneous in their preferences and the 1 provider performance can be enhanced relative weight they place on these sources Department of Health Policy, London School of Economics, London, UK by nurturing the altruistic motivation of motivation. 2Centro de Investigación en Salud Materna e Infantil of providers. The first comes from the When designing motivational and Centro de Investigación para el Desarrollo Integral US, where healthcare professionals in a instruments, policy makers also need to y Sostenible, Universidad Peruana Cayetano Heredia, hospital were more likely to change their take into account the broader environment. Lima, Peru behaviour and adopt better practices when Incentives usually target one aspect of 3Department of Health Policy and Management, Harvard a campaign for hand washing highlighted behaviour, but they are introduced into T H Chan School of Public Health, Boston, MA, USA the benefits for patients rather than the a complex system with an existing set of Correspondence to: I Papanicolas [email protected] benefit to themselves.60 The second comes cultures and constraints that may affect from a low income setting, where altruistic the willingness and ability of practitioners motivation might be expected to have to respond as intended. This makes it limited effects given the low remuneration difficult to predict the outcome of any This is an Open Access article distributed in and challenging working conditions. Yet, intervention, or to generalise about accordance with the Creative Commons Attribution in a recent experiment in India, community the transferability of findings across Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build health workers who received regular systems. For example, in settings where upon this work non-commercially, and license information highlighting the benefits they governance is weak, or political will their derivative works on different terms, provided were creating for patients increased their limited, incentives to improve provider the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/ performance by 25%, and this effect was performance have failed. Several studies licenses/by-nc/4.0/. strongest for those with high levels of have described incentive programmes intrinsic and altruistic motivation.61 These that failed to reduce absenteeism because examples highlight the need to identify of officials’ reluctance to fully implement 65 66 policies that can preserve and nurture such the monitoring or incentive systems, 1 Kruk ME, Larson E, Twum-Danso NA. Time for a reservoirs of goodwill. or politicians interfering with bureaucrat quality revolution in global health. Lancet Glob sanctions.67 This does not mean that Health 2016;4:e594-6. doi:10.1016/S2214- 109X(16)30131-0 Improving provider motivation incentives cannot work in some settings but 2 Mohanan M, Vera-Hernández M, Das V, et al. Not everyone is motivated in the same that they require innovative solutions, such The know-do gap in quality of health care for way, 62 and the interventions that we have as finding trusted entities to hold providers childhood diarrhea and pneumonia in rural India. JAMA Pediatr 2015;169:349-57. doi:10.1001/ described should not be seen as mutu- accountable, such as peers or the broader jamapediatrics.2014.3445 ally exclusive. Instead they form a palette community. 3 Das J, Holla A, Mohpal A, Muralidharan K. Quality of options from which policy makers can Technology could also provide new and accountability in healthcare delivery: audit- opportunities to address these challenges. study evidence from primary care in India. Am choose to design the most relevant solu- Econ Rev 2016;106:3765-99. doi:10.1257/ tion. This process should start with a needs Mobile technologies provide cheap and aer.20151138 assessment to identify and understand the flexible solutions to improve information 4 Das J, Hammer J, Leonard K. The quality of medical local obstacles that limit performance and systems and feedback, support clinical advice in low-income countries. J Econ Perspect 2008;22:93-114. doi:10.1257/jep.22.2.93 undermine quality of care. In some set- decisions, facilitate and monitor delivery 5 National Academies of Sciences. Medicine. Crossing tings, preliminary problems might have of care in remote settings, and eliminate the global quality chasm: improving health care to be addressed before tackling low moti- many barriers to payment.68-70 Despite worldwide. National Academies Press, 2018. 6 Leonard KL, Masatu MC. Using the Hawthorne vation. For example, if staff do not have the enthusiasm around and potential of effect to examine the gap between a doctor’s the clinical skills to make a correct diag- these tools, there have been few credible best possible practice and actual performance. nosis, no amount of money or feedback attempts at evaluating their capabilities to J Dev Econ 2010;93:226-34. doi:10.1016/j. will increase technical quality of care. If support and increase provider motivation.61 jdeveco.2009.11.001 7 Willis-Shattuck M, Bidwell P, Thomas S, et al. essential basic drugs and equipment are Contributors and sources: ML has studied the Improving motivation and retention of health not available, incentives will not improve determinants of provider behaviour and quality of professionals in developing countries: a systematic the treatments provided to patients. care in low and middle income countries, IP has review. BMC Health Serv Res 2008;8:247. doi:10.1186/1472-6963-8-247 In other contexts, careful consideration studied policies for performance improvement in health systems, with a focus on high income 8 Hickam DH, Severance S, Feldstein A, et al. The effect should be given to the broader environment countries. LH has much experience in public health of health care working conditions on patient safety. in which health workers operate to research, particularly in child and maternal health, Evid Rep Technol Assess (Summ) 2003;74:1-3. understand their behaviour. Maslow’s health systems and human resources in health. 9 Aiken LH, Sloane DM, Bruyneel L, et al, RN4CAST 63 consortium. Nurse staffing and education seminal motivation theory states that Competing interests: We have read and understood and hospital mortality in nine European people seek to satisfy their most pressing BMJ policy on declaration of interests and have no countries: a retrospective observational study. physiological and safety needs before they relevant interests to declare. Lancet 2014;383:1824-30. doi:10.1016/S0140- 6736(13)62631-8 can be influenced by “growth” needs such Provenance and peer review: Commissioned; 10 Das J, Woskie L, Rajbhandari R, Abbasi K, Jha A. as reputational concerns or altruism. In externally peer reviewed. Rethinking assumptions about delivery of healthcare: other words, if their financial remuneration This article is part of a series commissioned by The implications for universal health coverage. BMJ 2018;361:k1716. doi:10.1136/bmj.k1716 is insufficient for them to make ends meet, BMJ based on an idea from the Harvard Global Health Institute. The BMJ retained full editorial control over 11 McGlynn EA, Asch SM, Adams J, et al. The quality of providers are likely to be predominantly external peer review, editing, and publication. Harvard health care delivered to adults in the United States. driven by financial motives. However, Global Health Institute paid the open access fees. N Engl J Med 2003;348:2635-45. doi:10.1056/ NEJMsa022615 satisfaction of needs is not an “all or Mylène Lagarde, associate professor of health 64 1 12 Scott A, Sivey P, Ait Ouakrim D, et al. The effect nothing” phenomenon. Similarly, we economics of financial incentives on the quality of health 2 have reinforced the importance of bearing Luis Huicho, researcher care provided by primary care physicians.

18 doi: 10.1136/bmj.l5210 | BMJ 2019;366:l5210 | the bmj UNIVERSAL HEALTH COVERAGE

Cochrane Database Syst Rev 2011;9:CD008451. reporting on clinical outcomes: a systematic the proportion of health professionals practising doi:10.1002/14651858.CD008451.pub2 review and meta-analysis. BMC Health Serv in rural and other underserved areas. Cochrane 13 Mendelson A, Kondo K, Damberg C, et al. The effects Res 2016;16:296. doi:10.1186/s12913-016- Database Syst Rev 2009;1:CD005314. of pay-for-performance programs on health, health 1543-y doi:10.1002/14651858.CD005314.pub2 care use, and processes of care: a systematic review. 30 Lindenauer PK, Remus D, Roman S, et al. Public 50 Lagarde M, Blaauw D, Cairns J. Cost-effectiveness Ann Intern Med 2017;166:341-53. doi:10.7326/ reporting and pay for performance in hospital quality analysis of human resources policy interventions M16-1881 improvement. N Engl J Med 2007;356:486-96. to address the shortage of nurses in rural South 14 Witter S, Fretheim A, Kessy FL, Lindahl AK. doi:10.1056/NEJMsa064964 Africa. Soc Sci Med 2012;75:801-6. doi:10.1016/j. Paying for performance to improve the delivery 31 Joynt KE, Orav EJ, Zheng J, Jha AK. Public reporting socscimed.2012.05.005 of health interventions in low- and middle- of mortality rates for hospitalized medicare patients 51 Patterson F, Cleland J, Cousans F. Selection methods in income countries. Cochrane Database Syst and trends in mortality for reported conditions. Ann healthcare professions: where are we now and where Rev 2012;2:CD007899. doi:10.1002/14651858. Intern Med 2016;165:153-60. doi:10.7326/M15- next?Adv Health Sci Educ Theory Pract 2017;22:229- CD007899.pub2 1462 42. doi:10.1007/s10459-017-9752-7 15 Sherry TB, Bauhoff S, Mohanan M. Multitasking 32 Ryan AM, Nallamothu BK, Dimick JB. Medicare’s 52 Patterson F, Prescott-Clements L, Zibarras L, Edwards and heterogeneous treatment effects in pay-for- public reporting initiative on hospital quality had H, Kerrin M, Cousans F. Recruiting for values in performance in health care: evidence from Rwanda. modest or no impact on mortality from three key healthcare: a preliminary review of the evidence. Am J Health Econ 2017;3:192-226. doi:10.1162/ conditions. Health Aff (Millwood) 2012;31:585-92. Adv Health Sci Educ Theory Pract 2016;21:859-81. AJHE_a_00072 doi:10.1377/hlthaff.2011.0719 doi:10.1007/s10459-014-9579-4 16 Basinga P, Gertler PJ, Binagwaho A, Soucat AL, Sturdy 33 Kolstad JT. Information and quality when motivation 53 Roy AD. Some thoughts on the distribution J, Vermeersch CM. Effect on maternal and child is intrinsic: evidence from surgeon report cards. of earnings. Oxf Econ Pap 1951;3:135-46. health services in Rwanda of payment to primary Am Econ Rev 2013;103:2875-910. doi:10.1257/ doi:10.1093/oxfordjournals.oep.a041827 health-care providers for performance: an impact aer.103.7.2875 54 Besley T, Ghatak M. Competition and incentives with evaluation. Lancet 2011;377:1421-8. doi:10.1016/ 34 Vallance AE, Fearnhead NS, Kuryba A, et al. Effect of motivated agents. Am Econ Rev 2005;95:616-36. S0140-6736(11)60177-3 public reporting of surgeons’ outcomes on patient doi:10.1257/0002828054201413 17 Gertler P, Vermeersch C. Using performance selection, “gaming,” and mortality in colorectal 55 Deserranno E. Financial incentives as signals: incentives to improve medical care productivity and cancer surgery in England: population based cohort experimental evidence from the recruitment health outcomes. NBER Working Paper No 19046. study. BMJ 2018;361:k1581. of village promoters in Uganda. Am Econ J National Bureau of Economic Research, 2013. 35 Blumenthal DM, Valsdottir LR, Zhao Y, et al. A Appl Econ 2019;11:277-317. doi:10.1257/ 18 Witter S, Fretheim A, Kessy FL, Lindahl AK. survey of interventional cardiologists’ attitudes app.20170670 Paying for performance to improve the delivery and beliefs about public reporting of percutaneous 56 Ashraf N, Bandiera O, Lee S. Losing prosociality in the of health interventions in low- and middle- coronary intervention. JAMA Cardiol 2018;3:629-34. quest for talent? Sorting, selection and productivity income countries. Cochrane Database Syst doi:10.1001/jamacardio.2018.1095 in the delivery of public services. Development Rev 2012;2:CD007899. doi:10.1002/14651858. 36 Dranove D, Kessler D, McClellan M, et al. Is more discussion papers 65. Suntory and Toyota CD007899.pub2 information better? the effects of ‘report cards’ on International Centres for Economics and Related 19 Currie J, Lin W, Meng J. Addressing antibiotic health care providers. J Polit Econ 2003;111:555-88. Disciplines, LSE, 2018. abuse in China: an experimental audit study. doi:10.1086/374180 57 Dal Bó E, Finan F, Rossi MA. Strengthening state J Dev Econ 2014;110:39-51. doi:10.1016/j. 37 World Bank. World development report 2004: capabilities: the role of financial incentives in the jdeveco.2014.05.006 making services work for poor people. World Bank, call to public service. Q J Econ 2013;128:1169-218. 20 Papanicolas I, McGuire A. Do financial incentives 2003. doi:10.1093/qje/qjt008 trump clinical guidance? Hip replacement in England 38 Bjorkman M, Svenson J. Power to the people: 58 Schouten LMT, Hulscher MEJL, van Everdingen JJ, and Scotland. J Health Econ 2015;44:25-36. evidence from a randomised field experiment Huijsman R, Grol RP. Evidence for the impact of doi:10.1016/j.jhealeco.2015.08.001 on community-based monitoring in Uganda. quality improvement collaboratives: systematic 21 Borghi J, Little R, Binyaruka P, Patouillard E, Q J Econ 2009;124:735-69. doi:10.1162/ review. BMJ 2008;336:1491-4. doi:10.1136/ Kuwawenaruwa A. In Tanzania, the many costs qjec.2009.124.2.735 bmj.39570.749884.BE of pay-for-performance leave open to debate 39 Devarajan S, Reinikka R. Making services work for 59 Larsson S, Lawyer P, Garellick G, Lindahl B, whether the strategy is cost-effective. Health poor people. World Bank, 2002. Lundström M. Use of 13 disease registries in Aff (Millwood) 2015;34:406-14. doi:10.1377/ 40 Austin JM, Jha AK, Romano PS, et al. National hospital 5 countries demonstrates the potential to use hlthaff.2014.0608 ratings systems share few common scores and outcome data to improve health care’s value. Health 22 Braithwaite J, Churruca K, Ellis LA, et al. Complexity may generate confusion instead of clarity. Health Aff (Millwood) 2012;31:220-7. doi:10.1377/ science in healthcare— aspirations, approaches, Aff (Millwood) 2015;34:423-30. doi:10.1377/ hlthaff.2011.0762 applications and accomplishments: a white paper. hlthaff.2014.0201 60 Grant AM, Hofmann DA. It’s not all about Australian Institute of Health Innovation, Macquarie 41 Delfgaauw J. Dedicated doctors: public and me: motivating hand hygiene among University, 2017. private provision of health care with altruistic health care professionals by focusing on 23 Eijkenaar F. Pay for performance in health physicians. Erasmus School of Economics, Erasmus patients. Psychol Sci 2011;22:1494-9. care: an international overview of initiatives. Universiteit Rotterdam. Tinbergen Institute, 2007. doi:10.1177/0956797611419172 Med Care Res Rev 2012;69:251-76. 42 Arrow K. Uncertainty and the welfare economics of 61 Lee S. Intrinsic incentives: a field experiment on doi:10.1177/1077558711432891 medical care. Am Econ Rev 1963;53:941-3. leveraging intrinsic motivation in public service 24 Van Herck P, De Smedt D, Annemans L, Remmen R, 43 Liu T. Ma C-tA. Health insurance, treatment plan, delivery. 2018. https://scholar.harvard.edu/files/lee/ Rosenthal MB, Sermeus W. Systematic review: Effects, and delegation to altruistic physician. J Econ files/scott_lee_intrinsic_incentives_latest.pdf design choices, and context of pay-for-performance Behav Organ 2013;85:79-96. doi:10.1016/j. 62 Herzberg F. One more time: how do you motivate in health care. BMC Health Serv Res 2010;10:247. jebo.2012.11.002 employees? 1968. Harv Bus Rev 2003;81:87-96. doi:10.1186/1472-6963-10-247 44 Deci E, Ryan R. Intrinsic motivation and self- 63 Maslow AH. A theory of human motivation. Psychol 25 Emanuel EJ, Ubel PA, Kessler JB, et al. USing determination in human behavior. Springer, 1985. Rev 1943;50:370-96. doi:10.1037/h0054346 behavioral economics to design physician doi:10.1007/978-1-4899-2271-7 64 Maslow A. Motivation and personality. 3rd ed. incentives that deliver high-value care. Ann Intern 45 Patterson F, Knight A, Dowell J, Nicholson S, Cousans Pearson Education, 1987. Med 2016;164:114-9. doi:10.7326/M15-1330 F, Cleland J. How effective are selection methods 65 Dhaliwal I, Hanna R. The devil is in the details: the 26 Donato K, Miller G, Mohanan M, Truskinovsky in medical education? A systematic review. Med successes and limitations of bureaucratic reform in Y, Vera-Hernandez M. Personality Traits and Educ 2016;50:36-60. doi:10.1111/medu.12817 India. J Dev Econ 2017;124:1-21. doi:10.1016/j. Performance Contracts: Evidence from a Field 46 Lagarde M, Blaauw D. Pro-social preferences and jdeveco.2016.08.008 Experiment among Maternity Care Providers in India. self-selection into jobs: Evidence from South African 66 Banerjee AV, Glennerster R, Duflo E. Putting Am Econ Rev 2017;107:506-10. doi:10.1257/aer. nurses. J Econ Behav Organ 2014;107:136-52. a band-aid on a corpse: incentives for nurses p20171105 doi:10.1016/j.jebo.2014.09.004. in the indian public health care system. J Eur 27 Frey B. Not just for the money, an economic theory of 47 Brock JM, Lange A, Leonard KL. Generosity norms Econ Assoc 2008;6:487-500. doi:10.1162/ personal motivation. Edward Elgar, 1997. and intrinsic motivation in health care provision: JEEA.2008.6.2-3.487 28 Berwick DM, James B, Coye MJ. Connections evidence from the laboratory and field. J Hum 67 Callen M, Gulzar S, Hasanain SA, et al. The political between quality measurement and Resour 2016;51:133-62. doi:10.3368/jhr.51.1.133 economy of public sector absence: experimental improvement. Med Care 2003;41(Suppl):I30-8. 48 Ashraf N, Bandiera O. Altruistic capital. Am Econ evidence from Pakistan. National Bureau of Economic doi:10.1097/00005650-200301001-00004 Rev 2017;107:70-5. doi:10.1257/aer.p20171097 Research Working Paper No 22340. NBER, 2016. 29 Campanella P, Vukovic V, Parente P, Sulejmani A, 49 Grobler L, Marais BJ, Mabunda SA, Marindi PN, 68 White A, Thomas DS, Ezeanochie N, Bull S. Ricciardi W, Specchia ML. The impact of public Reuter H, Volmink J. Interventions for increasing Health worker mhealth utilization: a systematic the bmj | BMJ 2019;366:l5210 | doi: 10.1136/bmj.l5210 19 UNIVERSAL HEALTH COVERAGE

review. Comput Inform Nurs 2016;34:206-13. Innovations. Bull World Health Organ 2012;90:332- visual framework. Glob Health Sci Pract 2013;1:160- doi:10.1097/CIN.0000000000000231 40. doi:10.2471/BLT.11.099820 71. doi:10.9745/GHSP-D-13-00031 69 Lewis T, Synowiec C, Lagomarsino G, Schweitzer 70 Labrique AB, Vasudevan L, Kochi E, Fabricant R, J. E-health in low- and middle-income countries: Mehl G. mHealth innovations as health system Cite this as: BMJ 2019;366:l5210 findings from the Center for Health Market strengthening tools: 12 common applications and a http://dx.doi.org/10.1136/bmj.l5210

20 doi: 10.1136/bmj.l5210 | BMJ 2019;366:l5210 | the bmj UNIVERSAL HEALTH COVERAGE

Overcoming distrust to deliver universal health coverage: lessons from Ebola Liana Woskie and Mosoka Fallah use the Ebola outbreak in Liberia to better understand the role and consequences of distrust in health systems and how it affects universal health coverage

pidemics of infectious disease obtain needed services for themselves or little reason for West Point residents to trust often highlight underlying weak- their children if they became sick.4 the system when told about a strange new nesses in health systems. The Low rates of early care seeking are disease that required strict isolation from two most recent outbreaks of thought to have increased mortality from their families. Ebola virus disease, for example, Ebola. But early presentation is also Eexposed high levels of distrust that contrib- fundamental to mitigating unnecessary Care seeking and cooperation uted to the spread of disease but also have morbidity and mortality associated with Distrust in government (including govern- implications for universal health coverage. diseases from diabetes to HIV/AIDs. We ment provided healthcare) and exposure By the end of August 2019 just over a quar- know surprisingly little about the state of to negative Ebola related experiences were ter of deaths from Ebola in the Democratic health system distrust or what drives it. among the most important determinants of Republic of the Congo (DRC) had occurred We use the Ebola outbreak in West Point, care seeking in Monrovia, Liberia, towards 10 outside treatment centres.1 Since the treat- Liberia (the largest slum in the country’s the end of the Ebola outbreak. In West ment protocol includes isolation, this capital city) to illustrate how distrust in the Point this also extended to life saving medi- suggests that people were refraining from health system undermined care coverage cal advice, such as reporting of deaths and seeking care when symptoms arise or not when it was most needed and lay out three comprehensive contact tracing (box 2). remaining in treatment for the suggested strategies to better understand and tackle More recently, a population based duration. distrust within the broader context of study in the DRC identified low trust in One reason for this is lack of trust universal health coverage (UHC). institutions and belief as being associated in institutions, and specifically health with a decreased likelihood of adopting Distrust in the context of UHC preventive behaviours, including systems. Surveys conducted in North 11 Kivu, the centre of the outbreak, during The goal of UHC is to ensure that the whole acceptance of Ebola vaccines. Similar late 2018 to early 2019 found that people population, including the most disadvan- findings were reported in a survey of other African countries: “A staggering proportion viewed Ebola as a government scheme to taged groups, receives essential health services that are good quality. Tanehashi’s of citizens in most of the sampled countries marginalise certain groups or as part of a 1978 framework for assessing healthcare reported having gone without medicines business to profit aid workers, researchers, coverage sets out five stages from available or medical treatment in the previous and government officials.2 These findings to effective (fig 1).5 6 Health system distrust year, and going without health care was parallel those of a similar study conducted is a mediating factor that may drive down in Liberia during the west African Ebola the willingness of people to use health ser- Potential coverage crisis in 2014-15.3 In Liberia the distrust vices (“acceptability coverage”). If people was evident before the crisis, with Availability coverage find health services to be unacceptable and “People for whom service is available” another survey finding that about half of are unwilling to use them, they may remain respondents did not believe that they could uncovered even if services are technically in place.5 Accessibility coverage KEY MESSAGES “People who can use the service” Consequences of distrust • Acute disease outbreaks often shed Health system distrust light on underlying health system In August 2014, the transmission of Ebola failures in West Point was seen as a potentially High rates of distrust health system insurmountable threat to containing the Acceptability coverage • 7 “People who are willing to use the service” distrust have been exposed in both disease. The combined historical chal- recent Ebola epidemics. lenges of marginalisation, poor public health infrastructure, and poor healthcare Actual coverage • Health system distrust can make peo- ple less willing to use health services had resulted in residents of West Point seri- Contact coverage ‘‘People who use the service” in both acute and non-acute situations ously doubting the health system (box 1). A lack of clear expectations, To build acceptability countries must • miscommunication regarding what should routinely assess rates of distrust and Effective coverage be expected from the health system, and an its drivers, encourage efforts that ‘‘People who receive effective care” inability to deliver quality services under build confidence, and trust patients’ earlier health schemes set a challenging decision making processes Fig 1 | Tanehashi’s stages of healthcare baseline. When Ebola arrived, there was service coverage the bmj | BMJ 2019;366:l5482 | doi: 10.1136/bmj.l5482 21 UNIVERSAL HEALTH COVERAGE

presentation greatly affects chances of Box 1: Health system context in West Point, Liberia before Ebola survival, and for other acute and chronic Before the 2014 Ebola outbreak and after the Liberian civil war, West Point was known as conditions. Acceptance of and adherence a strong political base of George Weah, leader of the then opposition party Congress for to antiretrovirals, for example, have been Democratic Change. Residents of West Point felt that the government was not operating in found to be significantly associated with their interests because of their political support for Weah and a history of low social service trust in medications, trust in the healthcare provision. The area had inadequate refuse collection, sewage infrastructure, and latrines. The system, and a patient’s relationships with population (about 80 000 residents) was served by only one health centre, a joint government physicians and peers.14 and Catholic run clinic that provided free care. The Liberian Ministry of Health and Social Welfare (MOHSW) developed its first national Informal care seeking health policy and plan in 2007, which was centred on a basic package of health services. The A 2008 study in Liberia found that low con- policy was rolled out in about 80% of health facilities. However, communication about what 8 9 fidence in the government was correlated was covered was unclear. The availability of services increased under the scheme, but the with a greater reliance on the informal experience of residents of West Point was mixed. Although all government health facilities healthcare sector.4 In addition, people were were meant to be free, a survey in 2014 by the Community-Based Initiative, an organisation less likely to report confidence in the health set up to engage the community in tackling Ebola, found that many community members were system if they were in the lowest wealth paying large sums out of pocket. In addition, when people sought care, the clinic was often group.4 Earlier research found that informal unable to meet their needs—for example, drugs to treat postpartum haemorrhage were often healthcare visits in Liberia decreased with not available. Although not well documented, data from the national demographic and health a person’s wealth and satisfaction with the survey suggested that maternal mortality had risen from a baseline of about 770/100 000 formal health system.15 These factors may births to 970/100 000 in 2013. The poor service delivery was compounded by concerns that be related—wealthier people may get bet- people would not be treated with respect when they did seek care. ter care, be more satisfied with that care, and, in turn, be more trusting of the health most strongly correlated with views on has been associated with underuse of 12 13 system. Regardless, high rates of informal health services.” Distrust in the health recommended preventive services. This care seeking can be challenging for health system, and government more broadly, is relevant for Ebola, where timing of ministries working to achieve UHC. How many people receive their care through Box 2: Consequences of distrust in West Point—a personal account* informal sources, and how good that care During the Ebola outbreak I worked with the Community-Based Initiative (CBI), an organisation is, is difficult to quantify. As such, it is rarely started to mitigate distrust and mobilise communities in the fight against Ebola. In August accounted for in assessments of coverage.16 2018 the CBI discovered that secret burials were taking place in West Point, and I alerted the In addition, when informal care seeking WHO representative in Liberia, Nestor Ndayimirije. He proposed a secret meeting outside West becomes normal, it is difficult to change Point to gather information and protect people who might provide us with valuable insight into this behaviour quickly in situations of what was going on. These people met us at a private location and confirmed the secret burials. population risk and acute individual As a result we moved in with the burial team and picked up nine corpses. need. Such situations require a centralised We then arranged a meeting with community leaders, youth leaders, and women’s leaders. strategy to communicate to health They told us they did not think Ebola was in West Point. We asked them about the nine bodies providers, coordinate care, and ensure a we had taken in one day. An elder responded that nine deaths in a day is normal. high level of quality. Although a population Our inability to initiate basic public health measures to reduce the disease burden among may have “contact coverage” or, in some people in West Point who had major sanitation problems meant that death was normal. Why cases, even “effectiveness coverage” should they believe that these new deaths were due to a new phenomenon called Ebola? through the informal sector, high rates When Ebola services were introduced in West Point, many people interpreted these efforts, of informal care seeking therefore pose such as holding centres, as a government strategy to introduce Ebola to the population challenges to UHC. because of its political views. Building trust Harm to the health system Perceptions about healthcare in the Ebola treatment unit began to shift in late October. We In extreme instances, high levels of dis- had worked on community engagement in West Point through the Community-Based Initiative trust may threaten health providers. For (CBI), an organisation we started to mitigate distrust and mobilise communities in the fight example, the city of Butembo in DRC saw against Ebola. As people recognised that Ebola was a real threat, there was some reversal armed assaults on Ebola treatment centres, of the distrust emanating from West Point, which had previously led to the ransacking of the the murder of a WHO doctor, and frequent 17 centre used to treat residents with Ebola. attacks on Ebola vaccination teams. These However, as late as mid-October, we found people were still hiding corpses and secretly attacks may have been motivated by mis- taking them out of West Point for burial. We decided to hold a focus group discussion with the information, but high levels of underlying elders and community leaders. One of the key reasons people provided for secrecy around distrust in the health system seem to be an 17 burial was rumours that no one ever returned from the Ebola treatment unit alive. They went important factor. Responders were forced on to inform us that they were told that when relatives went to the unit they were killed, after to pause activities such as active case find- which their heads were cracked open and their bodies burnt without anyone informing their ing, contact tracing, and even the admin- loved ones. One of the local chiefs looked at me and said, “Dr Fallah, I won’t lie to you, if my istration of vaccines. The attacks affected relative is sick or dying in West Point, I will run away with him instead of taking him to that who was willing to work in the area as well unit.” as the costs of providing health services. *Reflections from Mosoka P Fallah based on work conducted with the Community-Based Initiative (CBI) that was founded in Similar problems affect other countries 2014 to shift Ebola transmission dynamics and funded by the United Nations Development Programme. that are striving to ensure universal

22 doi: 10.1136/bmj.l5482 | BMJ 2019;366:l5482 | the bmj UNIVERSAL HEALTH COVERAGE

coverage. Violence against doctors in both of trust within the health system.24 Among quality care across low and middle income east and south Asia, for example, seems to validated scales, the group-based medi- countries,34 accounting for between 5.7 have increased over the past 10 years with cal mistrust scale, medical mistrust index, and 8.4 million deaths a year.35 We lack a doctors in India, China, Pakistan, Nepal, and healthcare system distrust scale were similar quantification for patient centred and Sri Lanka all stating concern for their most commonly used.25 Table 1 give some care, but studies indicate that disrespect physical safety.18 The population’s lack examples of the questions and the different and abusive treatment of patients is com- of trust in medical institutions has been contexts in which they were applied. mon.36 A study in Liberia found that people suggested to be a driver of this violence.18 Given the prevalence of distrust in with low confidence in the health system In China, physicians have reported high historically marginalised populations, it were more likely to have been dissatis- exposure to verbal abuse, threats of assault, is important to thoughtfully adapt these fied with their last health visit.4 Traumatic and physical assaults, leading to emotional tools to new contexts and disaggregate data experiences during Ebola treatment were exhaustion and lower job satisfaction, with to identify what is driving that distrust. also found to be associated with distrust.37 many intending to leave their role.19 Nurses Patient level factors that drive distrust are Earlier work, such as a 2011 analysis of are also affected, with 7.8% of nurses in a not wholly predictable or consistent across citizens’ perceptions of health systems in 2015 Chinese study reporting physical contexts. Race may be an important factor 20 sub-Saharan African countries found violence and 71.9% reporting non-physical in informing health system distrust in the that quality of care was strongly associ- abuse in the preceding year.20 Most US whereas caste or religion may be more ated with public opinion of the overall perpetrators were patients or their relatives. relevant in India. Patient level factors can health system.38 There are limitations on As news of these events spreads, there is even have varying within countries. For how accurately patients can evaluate qual- concern that they may breed more fear and example, in China a population based study ity of care, but receipt of poor quality care insecurity and contribute to further loss of found that high education tracked with and patients’ experiences of care seem to confidence in the health system.21 high distrust, but another study among inform health system distrust. people who had received care in Shanghai Improving quality of care offers potential Looking forward: a health systems approach hospitals found that more education was to counter distrust. For example, in Many drivers of distrust in public institu- correlated with more trust.32 33 It also Liberia, once people who survived Ebola tions lie outside the purview of the health matters who measures distrust and how. returned to their communities shared their system, such as weak state capacity or his- It can be particularly problematic when a experiences, perceptions of the healthcare tory of civil unrest and war.22 It is logical distrusted group (eg, government actors system shifted (box 3). that populations faced with geographical in a fragile state) is associated with data Simply put, health systems may need to constraint who are poor and consistently collection. prove their worth more actively. This can neglected by the public sector will have lim- Regardless, we lack a comprehensive be done in various ways, such as providing ited trust in government. However, below picture of what drives distrust in countries incentives for known drivers of trust, we focus on historical betrayals of trust that are working to reach UHC. The drivers including provision of correct and safe care committed by or within the health system of distrust are diverse—people may doubt and ensuring positive patient experiences. that could be targeted to help reach univer- the integrity of a ruling party or may have Highlighting success and improving sal coverage. been harmed when seeking healthcare in transparency are also important. the past—and require different strategies. Routinely assess rates of distrust and drivers We currently lack the data to disentangle Trust patients and engage in true partnership Pandemic risk models have begun to quan- distrust and strategically address the For populations with low trust in the for- tify the effect of non-epidemiological fac- problem. mal sector, it is critical to understand tors on disease spread.23 Efforts to assess what patients do trust and why. Informal the potential effect of UHC investments may Encourage efforts to build underlying providers may be more readily acces- benefit from a similar strategy. Although confidence sible to rural populations or people liv- trust is often considered a qualitative con- Three global reports released in 2018 ing in slums, and they may also be more cept, we do have methods to routinely broadly defined high quality care as safe, “acceptable” because of concerns about assess it. In 2013, a systematic review of effective, and patient centred; the reports disrespect, abuse, or poor quality services. scales and indices identified 45 measures highlighted strikingly high rates of poor In some contexts, care provided in the

Table 1 | Example questionnaire items that assess aspects of trust in health system24 Example questionnaire items Surveyed population Object of trust “Despite my unfamiliarity with doctors, nurses, and hospitals, I feel Elderly US population with chronic disease26 Treatment Confidence in very confident about my treatment.” service “If you or your child is very sick tomorrow, can you get the health care General household sample, rural Liberia4 Healthcare system you need?” “Patients receive high quality medical care from the Health Care African American general sample, ­Philadelphia, USA27 Healthcare system System” Competence “I think my doctor may not refer me to a specialist when needed” General national population,USA28 Physicians “How well is the government doing in providing health care?” General household sample, rural Liberia4 Government “If a mistake were made in my health care, the health care system General population (jurors waiting at Municipal Court Healthcare system would try to hide it from me.” of Philadelphia)29 Honesty and Random sample of residents with heart conditions in “Patients have sometimes been deceived or misled at hospitals” Hospitals integrity Baltimore City, USA30 “Medical decisions are influenced by how much money [my provider] General population: villagers with and without insur- Healthcare providers can make” ance, Cambodia31 the bmj | BMJ 2019;366:l5482 | doi: 10.1136/bmj.l5482 23 UNIVERSAL HEALTH COVERAGE

1 Unicef. DRC Ebola situation report, 2 Sep 2019. Box 3: Community input to counter distrust* https://www.unicef.org/appeals/files/UNICEF_DRC_ By October 2014 West Point residents had began to understand that Ebola was real, but Humanitarian_SitRep_Ebola_2_Sept_2019.pdf 2 Building trust is essential to combat the Ebola distrust in the system persisted and people still did not go to the Ebola treatment unit when outbreak. Nature 2019;567:433-433. doi:10.1038/ they had symptoms. As Ebola spread, the Community-Based Initiative was faced with a serious d41586-019-00892-6 dilemma and ran the risk of undermining the trust that we had built over the past two months. 3 Blair RA, Morse BS, Tsai LL. Public health and public We asked residents of West Point whether showing them people from their community who trust: Survey evidence from the Ebola Virus Disease epidemic in Liberia. Soc Sci Med 2017;172:89-97. had survived Ebola survivors would change their minds about the treatment unit, and they said doi:10.1016/j.socscimed.2016.11.016 it would. 4 Svoronos T, Macauley RJ, Kruk ME. Can the The following week, West Point organised a large town hall meeting with local leaders, youth, health system deliver? Determinants of rural Liberians’ confidence in health care. Health Policy women, and children. Eleven Ebola survivors from West Point shared their experience and the Plan 2015;30:823-9. doi:10.1093/heapol/czu065 role that treatment played in their survival. 5 Hogan DR, Stevens GA, Hosseinpoor AR, Boerma The chief who had previously told me he would run away with his relatives instead of going T. Monitoring universal health coverage within the Sustainable Development Goals: development to the treatment unit (box 2) turned to us and said, “Now I see with my own eyes and believe in and baseline data for an index of essential health the unit.” By the end of October, we had moved 28 patients with Ebola to the unit by working services. Lancet Glob Health 2018;6:e152-68. with the elders and chiefs. They were the last group of confirmed Ebola cases in West Point. For doi:10.1016/S2214-109X(17)30472-2 us, this reinforced that access to high quality healthcare with visible results has the propensity 6 Tanahashi T. Health service coverage and its evaluation. Bull World Health Organ 1978;56:295- to shift distrust in the health system. 303. https://www.ncbi.nlm.nih.gov/pmc/articles/ *Reflections from Mosoka P Fallah based on work conducted with the Community-Based Initiative (CBI) that was founded in PMC2395571/pdf/bullwho00439-0136.pdf. 2014 to shift Ebola transmission dynamics and funded by the United Nations Development Programme. 7 Fallah M, Dahn B, Nyenswah TG, et al. Interrupting Ebola transmission in Liberia through community- based initiatives. Ann Intern Med 2016;164:367-9. informal sector may even be of compara- paper. MPF reviewed, edited, and oversaw the paper. doi:10.7326/M15-1464 ble standard to that in the formal sector.39 MPF led Ebola community based initiatives in West 8 Petit D, Sondorp E, Mayhew S, Roura M, Roberts B. Point slum, where he grew up. These activities were Implementing a basic package of health services in It is important not to assume that patients run through the Community-Based Initiative that was post-conflict Liberia: perceptions of key stakeholders. are naive in their assessment of, and corre- founded in 2014 to shift Ebola transmission dynamics Soc Sci Med 2013;78:42-9. doi:10.1016/j. sponding choices in, healthcare. Strategies and funded by the United Nations Development socscimed.2012.11.026 Programme. Some aspects of this work have been 9 Government of Liberia. Ministry of Health and Social that treat communities themselves as the described previously.7 LRW worked on the team Welfare. Community health survey for health seeking primary barrier to ensuring care coverage commissioned by the National Academy of Medicine behaviour and health financing in Liberia. Monrovia, (eg, through behaviour change) may lose to generate data for its evaluation of low and middle 2008. sight of fundamental problems with the income country health system quality in Crossing the 10 Morse B, Grépin KA, Blair RA, Tsai L. Patterns Global Quality Chasm. of demand for non-Ebola health services health system to which the population is during and after the Ebola outbreak: panel responding. Competing interests: We have read and understood survey evidence from Monrovia, Liberia. BMJ BMJ policy on declaration of interests and have no It is critical to engage with communities— Glob Health 2016;1:e000007. doi:10.1136/ relevant interests to declare. bmjgh-2015-000007 not just educate and inform. This should 11 Vinck P, Pham PN, Bindu KK, Bedford J, Nilles include an honest assessment of where Provenance and peer review: Commissioned; EJ. Institutional trust and misinformation in the externally peer reviewed. people choose to seek care and why. response to the 2018-19 Ebola outbreak in North This article is part of a series commissioned by The It should take population concerns Kivu, DR Congo: a population-based survey. Lancet BMJ based on an idea from the Harvard Global Health Infect Dis 2019;19:529-36. doi:10.1016/S1473- seriously and be guided by those who Institute. The BMJ retained full editorial control over 3099(19)30063-5 express distrust in the health system. external peer review, editing, and publication. Harvard 12 Abiola SE, Gonzales R, Blendon RJ, Benson J. This is challenging in conflict situations, Global Health Institute paid the open access fees. Survey in sub-Saharan Africa shows substantial Liana R Woskie, research fellow1,2 support for government efforts to improve health which also highlights the need to more services. Health Aff (Millwood) 2011;30:1478-87. Mosoka P Fallah, deputy director general for systematically capture strategies that work. doi:10.1377/hlthaff.2010.1055 technical services3,4 13 O’Malley AS, Sheppard VB, Schwartz M, Mandelblatt 1Harvard Global Health Institute, Cambridge, MA, USA Conclusion J. The role of trust in use of preventive services 2Department of Health Policy, London School of among low-income African-American women. Ebola provides a stark example, but dis- Economics and Political Science, London, UK Prev Med 2004;38:777-85. doi:10.1016/j. ypmed.2004.01.018 trust undermines investments in UHC 3National Public Health Institute of Liberia, Monrovia, across the care continuum. Health system 14 Mostashari F, Riley E, Selwyn PA, Altice FL. Acceptance Liberia and adherence with antiretroviral therapy among 4 distrust is not fully understood but seems to Harvard Medical School, Boston, MA, USA HIV-infected women in a correctional facility. J Acquir be partly driven by the health system itself; Correspondence to: M P Fallah Immune Defic Syndr Hum Retrovirol 1998;18:341-8. it is both historically grounded and highly [email protected] doi:10.1097/00042560-199808010-00005 15 Kruk ME, Rockers PC, Varpilah ST, Macauley R. rational. This should raise concern, but it Which doctor?: Determinants of utilization of also provides cause for optimism. We must formal and informal health care in postconflict act on the modifiable causes of distrust if liberia. Med Care 2011;49:585-91. doi:10.1097/ This is an Open Access article distributed in MLR.0b013e31820f0dd4 we want to deliver on the promise of UHC, accordance with the Creative Commons Attribution 16 Hogan DR, Stevens GA, Hosseinpoor AR, Boerma providing not just superficial coverage but Non Commercial (CC BY-NC 4.0) license, which T. Monitoring universal health coverage within the the high quality healthcare that people permits others to distribute, remix, adapt, build Sustainable Development Goals: development upon this work non-commercially, and license want. and baseline data for an index of essential health their derivative works on different terms, provided services. Lancet Glob Health 2018;6:e152-68. the original work is properly cited and the use is Contributors and sources: LRW and MPF worked doi:10.1016/S2214-109X(17)30472-2 non-commercial. See: http://creativecommons.org/ on the Harvard-LSHTM Independent Panel on the 17 Beaumont P. “Most complex health crisis in history”: licenses/by-nc/4.0/. Global Response to Ebola in 2015 and 2016 as well Congo struggles to contain Ebola. Guardian 2019 as related work with the ministries of health from the Jun 25. https://www.theguardian.com/global- three most affected countries. Building on this, LRW development/2019/jun/25/most-complex-health- and MPF conceived of the paper. LRW drafted the crisis-congo-struggles-ebola-drc

24 doi: 10.1136/bmj.l5482 | BMJ 2019;366:l5482 | the bmj UNIVERSAL HEALTH COVERAGE

18 Ambesh P. Violence against doctors in the 26 Mascarenhas OA, Cardozo LJ, Afonso NM, et al. China. Chin Med J (Engl) 2016;129:814-8. Indian subcontinent: a rising bane. Indian Heart Hypothesized predictors of patient-physician trust doi:10.4103/0366-6999.178971 J 2016;68:749-50. doi:10.1016/j.ihj.2016.07.023 and distrust in the elderly: implications for health 34 Berwick DM, Kelley E, Kruk ME, Nishtar S, Pate MA. 19 Shi J, Wang S, Zhou P, et al. The frequency of patient- and disease management. Clin Interv Three global health-care quality reports in 2018. initiated violence and its psychological impact on Aging 2006;1:175-88. Lancet 2018;392:194-5. doi:10.1016/S0140- physicians in china: a cross-sectional study. PLoS 27 Shea JA, Micco E, Dean LT, McMurphy S, Schwartz 6736(18)31430-2 One 2015;10:e0128394. doi:10.1371/journal. JS, Armstrong K. Development of a revised 35 National Academy of Medicine. Crossing the global pone.0128394 health care system distrust scale. J Gen Intern quality chasm: improving health care worldwide. 20 Jiao M, Ning N, Li Y, et al. Workplace violence against Med 2008;23:727-32. doi:10.1007/s11606-008- 2018. http://www.nationalacademies.org/hmd/ nurses in Chinese hospitals: a cross-sectional survey. 0575-3 Reports/2018/crossing-global-quality-chasm- BMJ Open 2015;5:e006719-006719. doi:10.1136/ 28 Doescher MP, Saver BG, Franks P, Fiscella K. Racial improving-health-care-worldwide.aspx bmjopen-2014-006719 and ethnic disparities in perceptions of physician 36 Bowser D, Hill MPHK. Exploring evidence for 21 Wu D, Wang Y, Lam KF, Hesketh T. Health system style and trust. Arch Fam Med 2000;9:1156-63. disrespect and abuse in facility-based childbirth reforms, violence against doctors and job doi:10.1001/archfami.9.10.1156 report of a landscape analysis. 2010. https:// satisfaction in the medical profession: a cross- 29 Rose A, Peters N, Shea JA, Armstrong K. Development www.ghdonline.org/uploads/Respectful_Care_at_ sectional survey in Zhejiang Province, Eastern and testing of the health care system distrust scale. Birth_9-20-101_Final1.pdf China. BMJ Open 2014;4:e006431. doi:10.1136/ J Gen Intern Med 2004;19:57-63. doi:10.1111/ 37 Blair RA, Morse BS, Tsai LL. Public health and bmjopen-2014-006431 j.1525-1497.2004.21146.x public trust: Survey evidence from the Ebola 22 Dhillon RS, Kelly JD. Community trust and the 30 LaVeist TA, Isaac LA, Williams KP. Mistrust of Virus Disease epidemic in Liberia. Soc Sci Ebola endgame. N Engl J Med 2015;373:787-9. health care organizations is associated with Med 2017;172:89-97. doi:10.1016/j. doi:10.1056/NEJMp1508413 underutilization of health services. Health Serv socscimed.2016.11.016 23 Chowell G, Nishiura H. Transmission dynamics and Res 2009;44:2093-105. doi:10.1111/j.1475- 38 Abiola SE, Gonzales R, Blendon RJ, Benson J. control of Ebola virus disease (EVD): a review. BMC 6773.2009.01017.x Survey in sub-Saharan Africa shows substantial Med 2014;12:196. doi:10.1186/s12916-014- 31 Ozawa S, Walker DG. Comparison of trust in public support for government efforts to improve health 0196-0 vs private health care providers in rural Cambodia. services. Health Aff (Millwood) 2011;30:1478-87. 24 Ozawa S, Sripad P. How do you measure trust in the Health Policy Plan 2011;26(Suppl 1):i20-9. doi:10.1377/hlthaff.2010.1055 health system? A systematic review of the literature. doi:10.1093/heapol/czr045 39 Das J, Hammer J, Leonard K. The quality of Soc Sci Med 2013;91:10-4. doi:10.1016/j. 32 Duckett J, Hunt K, Munro N, Sutton M. Does medical advice in low-income countries. J Econ socscimed.2013.05.005 distrust in providers affect health-care utilization Perspect 2008;22:93-114. doi:10.1257/ 25 Williamson LD, Bigman CA. A systematic review in China?Health Policy Plan 2016;31:1001-9. jep.22.2.93 of medical mistrust measures. Patient Educ doi:10.1093/heapol/czw024 Couns 2018;101:1786-94. doi:10.1016/j. 33 Zhao D-H, Rao K-Q, Zhang Z-R. Patient trust in Cite this as: BMJ 2019;366:l5482 pec.2018.05.007 physicians: empirical evidence from Shanghai, http://dx.doi.org/10.1136/bmj.l5482

the bmj | BMJ 2019;366:l5482 | doi: 10.1136/bmj.l5482 25 UNIVERSAL HEALTH COVERAGE

Global health security and universal health coverage: from a marriage of convenience to a strategic, effective partnership

lobal health security (GHS) be in conflict. Similarly, while both agendas care services and access to safe, effective, and universal health coverage enshrine human rights and we see conver- quality essential medicines and vaccines (UHC) are frequently regarded gence through the realisation of the right for all”.7 as two sides of the same coin,1 to health, we see distinctions between eco- We define GHS activities as those or more cynically as a marriage nomic, cultural and social rights with civil concerned with preventing, detecting Gof convenience.2 Yet, there has been little and political rights. and responding to infectious disease consideration of how these ideals interact, It is important to address these differences threats of international concern to limit with academics and policymakers assum- before considering the mutual opportunities any socioeconomic impact of transborder ing that actions for one will also be advan- offered by their ‘marriage’, to ensure that disease, which mirrors the WHO tageous to the other. This paper analyses at inherent differences are not jettisoned for definition.8 Nevertheless, we recognise a macro level where these ideals converge, pragmatic reasons, risking distortion of that GHS is “very much like a chameleon” and where differences lie both conceptu- local health priorities. We support the link “essentially contested” and “not adequately ally and empirically. We argue both GHS that health system strengthening (HSS) defined”.9–12 and UHC focus on the mitigation of risk creates opportunity to connect GHS and GHS is underpinned by a legal and human rights. Mitigating the risk of UHC in a tangible way, with clear policy instrument, the International Health individuals who face impoverishment pathways that can benefit both ideals. Regulations (2005) (IHR).13 The IHR provide owing to healthcare expenditure is core to guidance for how states should develop UHC. For GHS, the risk is transnational and Defining Global Health Security (GHS) and and maintain their national capacities to emerges from outbreaks with cross-border Universal Health Coverage (UHC) minimise public health threats. While there potential. Hence, the bearer of the risk, and We recognise that the definition of UHC is no binding international legal equivalent the appropriate steps to mitigate it, are dif- can vary in distinct, but convergent ways.3 for UHC, the International Covenant on ferent in each agenda and may sometimes A holistic definition is ensuring individuals Economic, Social and Cultural Rights have access, without discrimination to com- guarantees the human right to health. SUMMARY BOX prehensive, appropriate and timely, quality General Comment 14 on the right to health, health services determined at the national which provides interpretive guidance on What is already known about this ­subject? level according to needs, as well as access the right to health, proposes a framework • Universal health coverage (UHC) and to safe and affordable medicines, while of availability, accessibility, acceptability global health security (GHS) are fre- ensuring that the use of these services do and quality.14 Moreover, policy initiatives quently being used in tandem by not expose users to financial difficulties.4 create normative guidance on how to policymakers, recognising that there However, for the purpose of this paper, we implement UHC, including The World are synergies between the two paral- focus within this definition on the extent to Health Report 2010,5 Making Fair Choices lel agendas. which the costs of healthcare are covered.5 on the Path to UHC and the United Nations What are the new findings We recognise that such a definition is not General Assembly 67/81.15 Similarly, GHS • UHC and GHS goals are in tension. comprehensive, but we also acknowledge has the policy and operational work of the The research and practice communi- that the two components of UHC (access Global Health Security Agenda (GHSA), ties that represent these two streams and risk protection) are in tension when an international partnership launched need to engage so that smart strate- it comes to decision making about provi- in 2014 and now comprising over 60 gies can be identified to improve both sion, particularly in resource-poor settings, countries, international organisations 16 aims simultaneously using codepend- as the goal of access would lead to prioriti- and non-governmental stakeholders, ent, but distinct policy. sation of the most (cost) effective services, which provides political impetus and whereas a focus on financial protection • Risk and human rights are two areas of international collaboration to meeting IHR convergence between UHC and GHS. would favour allocation of resources to requirements. more expensive interventions.6 However, in Divergence appears in the concep - • stressing the importance of universal access tualisation of risk at the collective or Current intersections between the two to effective healthcare, and universal finan- concepts individual level, and the prioritisation cial protection against the costs of this care, Five key works have sought to connect GHS of domestic or global activity. the definition is consistent with the United and UHC. Jain and Alam highlight that UHC What are the recommendations for Nations Sustainable Development Goals can help advance GHS.17 First, low or no policy and practice? (SDGs), which includes in Goal 3 “ensure financial barriers to accessing healthcare • Health systems strengthening can be healthy lives and promote well-being for all stimulates demand for health services the policy mechanism which, brings at all ages” and in particular target (3.8) to which facilitates early infectious disease GHS and UHC together, elevating “achieve UHC including financial risk pro- detection. Second, protecting people from health and mitigating risk for all. tection, access to quality essential health- catastrophic financial risk reduces an indi-

26 Wenham C, et al. BMJ Glob Health 2019;4:e001145. doi:10.1136/bmjgh-2018-001145 UNIVERSAL HEALTH COVERAGE

vidual’s possibility of falling into poverty, al conflate UHC and HSS. These are used scale outbreak, threatening a population an important social determinant of infec- interchangeably, and this risks unclear and/or economic or political stability as tious disease spread.18 understandings of what UHC entails, a result of opportunity costs lost through Yates, Dhillon and Rannan-Eliya echo furthering the potential for misaligned interrupted access to international Jain and Alam’s first thesis verbalising priorities. markets, reduced international travel and “the availability of accessible and universal The fifth work considers GHS as fear among the population. Despite the healthcare services in all countries is the “collective” security and “individual” IHR seeking to minimise such disruption, crucial first line of defence for all against security which broadly aligns with there are several examples of factors such threats to health”.19 Moreover, if UHC.23 Heymann suggests that a beyond a government’s control during an people are unable to access healthcare in difference exists between collective health outbreak which impact a range of sectors their local communities, this increases the security concerned with mutual global beyond health.25–27 Indeed, President Ellen likelihood of individuals crossing borders vulnerabilities posed by transborder Johnson Sirleaf argued that the best action to seek healthcare, thus increasing the risk spread of acute public health issues, and the USA could take to support Liberia in of onward transmission internationally.19 individual health security which includes the Ebola epidemic was to “not ostracise This work shows how these two concepts access to safe and effective health services, us via trade”, suggesting that severing are mirrored empirically: suggesting that products and technologies.23 Heymann’s economic ties would pose as much risk as UHC’s relationship between financial argument follows that if there is individual the virus itself (Emily Mendenhall, personal protection and equitable distribution of health security, this contributes to communication, 2017). risk (which addresses people’s ability to collective health security at the community, Accordingly, GHS focuses on future- pay, while protecting the sick), mirrors the national and global levels (i.e GHS). proofing pandemic risk through relationship between donor and recipient preparedness. It does this by contingency states for GHS, whereby wealthy states Conceptual convergence: risk planning for a range of disease threats.28 finance outbreak responses in affected Both UHC and GHS aim to mitigate poten- Luckily, large-scale international outbreaks states. While this risk pooling is not part tial health and economic threats either are rare events, nevertheless, the severity of of the IHR or GHSA mandate, it can be at the level of the individual (UHC) or the the potential (socio)economic impact of an argued that IHR compliance should reflect collective (GHS). For UHC, one such risk outbreak leads to considerable investment the ability to pay while protecting weaker results from individuals’ exposure to eco- in risk mitigation. This inadvertently states.19 nomic hazard as a result of a health event, may bias the public’s risk perception, Moreover, Yates et al highlight that is, an individual’s health needs may creating potentially disruptive influences that movements towards UHC build be met only by incurring impoverishing or on “business as usual” for international trust.2 19 This form of trust may exist catastrophic costs associated with access- travel and trade.12 29 Exemplifying this between governments and populations, ing appropriate healthcare.19 This form of was the West-Africa Ebola epidemic, between health providers and patients individual or familial risk is centred on the which had a relatively low likelihood of and between financiers and recipients cost, rather than the type of illness and can ‘anyone in the globe’ becoming infected, of health. This trust may foster effective relate to acute to chronic conditions. Any- because of the low reproductive ratio of collaboration when an outbreak emerges, one may be exposed to this financial risk, the disease. Nevertheless, despite the low improving public compliance with the potential exposure is a lifetime, the actual risk, there was a high perceived state-led interventions to limit disease likelihood of occurrence is high, and the risk. Margaret Chan reflected “I have spread.20 21 However, Ooms et al are more consequences of exposure are dispropor- never seen a health event strike such sceptical of joining the two agendas tionately large for the poor who have insuf- fear and terror, well beyond the affected together, recognising that they are ficient funds to ensure financial resilience communities”.30 This fear led to the synergistic, but not self-evidently so.2 In when confronted with a health concern.24 implementation of expensive policies such resource-poor settings, they recognise However, UHC offers an effective risk as airport screening apparatus in HICs. distinct policy pathways for UHC and GHS; reduction intervention: proposing prepay- These were not instrumental in reducing for example, whether to fund development ment and pooling mechanisms to reduce the actual risk of disease incursion but were of surveillance capabilities or social health both the probability of healthcare-related effective political placebos implemented by insurance mechanisms, a point we would losses occurring, and the severity of their governments to reduce perceived risks felt agree with. impacts on household’s budgets when by HIC citizens. Ooms et al further underscore the they do. This also enhances individuals’ instrumentalism in linking these agendas. willingness and ability to access health- Conceptual convergence: human rights Tying UHC to GHS may provide greater care as opposed to delaying careseeking Heymann’s distinction between GHS as col- leverage for financing UHC2 (Yamey echoes until they become very ill, thereby driving lective security and UHC as individual secu- this suggestion, that while the world’s gaze up healthcare costs for everyone. Accord- rity allows convergence between the two is on GHS in the wake of Ebola, associating ingly, risk reduction through UHC benefits agendas through the lens of human rights these can be a tool for getting attention to both individuals and societies. Moreo- also. Achieving both GHS and UHC require UHC and the health of populations in low ver, reducing risk to any health concern states to comply with their obligations and middle-income (LMIC) settings22). through UHC, including communicable and duties under international, regional Conversely, GHS advocates may connect disease, has significant opportunity costs and domestic human rights law. Human with the UHC agenda to gain legitimacy for GHS. rights are often conceptualised as matters among those who conceive of the security Instead of the ‘livelihood risk’ for UHC, of individual security, whereby a state fails discourse being too focused on high income the risk for GHS results from an infectious to respect, protect or fulfil an individual’s country (HIC) interests.2 However, Ooms et disease hazard which may result in a large- human rights. However, even where an indi-

Wenham C, et al. BMJ Glob Health 2019;4:e001145. doi:10.1136/bmjgh-2018-001145 27 UNIVERSAL HEALTH COVERAGE

vidual successfully seeks recourse against case of UHC and GHS, as outlined above, yet focus on GHS when looking to attract a state for a human rights violation, such but so too is the ‘for whom’. For UHC, at risk donor dollars. decisions have a collective impact, setting is the everyday person who may be affected precedent that results in the state complying by ill health and the associated costs, or Practical convergence: HSS with its human rights obligations elsewhere. the inability to access health services due We argue that HSS can be the policy mecha- This is particularly the case for UHC, where to other non-financial barriers. For GHS, nism which brings GHS and UHC together, human rights actions launched by individ- however, the global population is at risk as elevating health and mitigating risk for all. uals have, according to some proponents, their chances of contracting an infectious This echoes Kutzin and Sparkes who argue, addressed underlying systemic failures by disease are reduced through ensuring “health system strengthening is what we governments to take steps to immediately GHS. Others have argued that the referent do: UHC, health security and resilience is realise the right to non-discrimination and object for GHS is the economy or national what we want”.3 progressively realise the right to health.31 security of a particular state fearing the A health system can be defined as These latter obligations typically fall within socioeconomic impact of an outbreak on the ensemble of all public and private the realm of economic, social and cultural trade and travel.35 36 Accordingly, GHS organisations, institutions and resources rights. This requires states to progressively predominantly mitigates risk from the top involved in the improvement, maintenance realise these rights to the maximum of their down, and UHC may mitigate risk from or restoration of health.37 HSS refers to available resources, while not regressing the bottom up, although infrastructure policy and programmatic activity designed from steps already taken for non-discrimi- and support is required from the state to apply systems thinking to health, to nation and meeting minimum core obliga- to support individuals in risk pooling improve overall performance.38 The WHO tions.32 In contrast, much of the dialogue behaviour. framework for HSS encapsulates six discussing GHS and human rights relates Both UHC and GHS risks are mitigated building blocks: service delivery, health to civil and political rights, such as those by financial investment in health. For UHC, workforce, health information systems, codified in the IHR; rights that the state the investment reduces the time people medical products, health financing and must respect, provide and fulfil such as the delay care-seeking due to the financial leadership and governance.39 The health rights to life, freedom of movement, and burden of paying for health. Through system shapes many people’s health by freedom from torture or cruel, inhuman or GHS, the investment is in pandemic determining how s/he accesses medical degrading treatment.33 While this civil and preparedness; strengthening surveillance care, from whom s/he receives medical political rights framing is understandable and response mechanisms to respond to care, what medicines are available as it focuses on the short-term and imme- infectious disease outbreaks under IHR and accessible, what technologies are diate vulnerability of individuals to the (2005). Consequentially, the rationales affordable and available for testing and state’s actions when seeking to protect the and methods for mitigating against these diagnostics and how s/he is expected to many and/or the economy during an out- risks—from the household to state levels— pay for it,39 and as such contains many of break, the goals of GHS are fundamentally are quite different. the tenets of UHC. grounded in economic, social and cultural While private and non-profit actors are For UHC, functioning health systems rights, namely, the right to health. The right vital in global health, we argue that states organised around people, institutions to health includes the obligation that states play a fundamental role in the convergence and resources leads to improved access, take steps necessary for the “prevention, of the two risks identified in this paper. quality, sustainability and affordability treatment and control epidemic, endemic, However, a distinction emerges between for individuals.38 For GHS, successfully occupational and other diseases”.34 This mitigating a state’s risks which are domestic functioning health systems underpin obligation is congruent with GHS, and is priorities, and those that are globally countries’ ability to detect and respond to also codified in the IHR, for example, within focused. For instance, state priorities that disease threats.39 In this way, a response the core capacity obligations. are domestically focused may involve to a health emergency (GHS) should be As a result, convergence between prepayment schemes to reduce the financial embedded within an existing health UHC and GHS can be found through the risks posed to citizens (UHC). On the other system, involving Farmer’s interweaving of realisation the right to health, with both hand, states prioritising GHS focus on “stuff, staff, systems and space” to address UHC and GHS requiring that states address implementing the IHR (2005) to reduce the the needs of an epidemic and population inaction or regression in realising the right risk of severe economic impact in the case health.40 Kluge expands this, providing to health to the mutual benefit of both of an acute public health event (Wenham, suggestions for how to interlink these ideals. Examining Sovereignty in Global Health, concepts, noting that investing in HSS PhD, 2016). These risks are fundamentally improves GHS, so that systems become Conceptual divergence: inward versus different, although the policies deployed resilient to health crises and can respond outward: individual versus global security may carry opportunity costs for both UHC when needed.41 By investing in health Despite unifying features, there are differ- and GHS goals. Governments, particularly systems, this increases the resilience of ences in each with respect to the charac- in resource-constrained settings must states to respond to outbreaks of disease terisation, who is identified as “at risk” and decide whether to prioritise their global or that spread across national borders, what responses have been taken to mitigate domestic responsibilities, based on which thereby investing indirectly in GHS.3 38 HSS risk. We suggest these understandings of risk they consider the most important. therefore is a common road to both UHC risk mirror divergent conceptualisations of National leaders may prioritise one agenda and GHS.3 security. over the other, aligning with political and GHS has sought to answer two questions: economic priorities; for example, they may Indicators convergence security from what and for whom?12 We prioritise UHC when fighting an election as Beyond the conceptual, we assessed con- know that the ‘from what’ is different in the it is popular with the domestic electorate, vergence and divergence of UHC, GHS

28 Wenham C, et al. BMJ Glob Health 2019;4:e001145. doi:10.1136/bmjgh-2018-001145 UNIVERSAL HEALTH COVERAGE

and HSS based on policy metrics. As these aim, where we see convergence is a direct GHS and UHC, there is considerable over- concepts are embedded within key pieces evaluation of how the concepts overlap. lap between HSS and both GHS and UHC, of global policy, it seemed appropriate to Figure 1 shows a tepid synergy between with each of the six building blocks finding use these indicators to ascertain whether UHC and GHS. Although UHC indicators a comparable indicator with the other two there was practical as well as conceptual explicitly include reference to GHS, in a agendas, and all three goals focusing on convergence between goals. We mapped catchall “Health Security IHR Core Capac- health workforce, access to medicines and GHS, using the first edition of the Joint ity Index”, it is not a key component of the financing/financial risk protection. External Evaluation Tool indicators as index. Convergence appeared in financing, a proxy, and UHC, using SDG indicators health workforce availability and capacity Concerns linking these agendas 3.8.1 and 3.8.2, to measure health service and access to medicines. There was not Synergising GHS and UHC raises several red coverage and financial protection24 42 43 and even overlap between the “infectious dis- flags. For UHC focusing on health through HSS, using the six WHO Building Blocks. ease” indicators of UHC and those of GHS. prepayment risks prioritising curative clini- As these indicators link to each policy However, despite limited overlap between cal services at the expense of individual

Figure 1 | Synergy between global health security (GHS), universal health coverage (UHC) and health systems strengthening (HSS).

Wenham C, et al. BMJ Glob Health 2019;4:e001145. doi:10.1136/bmjgh-2018-001145 29 UNIVERSAL HEALTH COVERAGE

and population health promotion and pre- promoting those parts of a health system that Received 30 August 2018 vention.44 This leads to more healthcare overlap between the two frameworks and Revised 19 October 2018 services but worse outcomes overall and overlooks what falls outside. Accordingly, we Accepted 28 October 2018 44 distributed benefits less equitably. For must consider what defines a strong health Clare Wenham,1 Rebecca Katz,2 Charles Birungi,3,4 GHS, the limitation is its focus on national system on an individual country basis that Lisa Boden,5 Mark Eccleston-Turner,6 and economic security and the threat of must address both the individual’s and the Lawrence Gostin,7 Renzo Guinto,8 Mark Hellowell,9 Kristine Husøy Onarheim,10 Joshua Hutton,11 infectious disease amid trade routes. This population’s needs. 12 13 Anuj Kapilashrami, Emily Mendenhall, prioritises diseases which affect dominant Alexandra Phelan,14 Marlee Tichenor,15 15 trading networks of HICs, creating a quasi- Conclusion Devi Sridhar 1 postcolonial power imbalance denoting UHC and GHS are increasingly linked in Department of Health Policy, London School of Economics and Political Science, London, UK which diseases are considered globally global health policy. This paper illumi- 2Center for Global Health Science and Security, important. nated the potential synergies between the There remain health issues which fall Georgetown University, Washington, District of Columbia, two parallel agendas, but has considered USA outside of both of GHS and UHC (and HSS) 3 the inherent tensions of a joined up UHC- Institute for Global Health, University College London, priority setting. Recognising the conditions GHS framework. We consider risk as being London, UK that are systematically excluded from both a unifying conceptual tool: the risk of the 4UNAIDS, Geneva, Switzerland agendas is equally important. For example, international spread of infectious disease 5Global Academy of Agriculture and FoodSecurity, The road safety, improvement of Water, on a population and national/economic Royal (Dick) School of Veterinary Studies and The Roslin Sanitation and Hygiene (WASH) facilities, security is fundamental to GHS. For UHC, Institute, University of Edinburgh, Edinburgh, UK pest control and neglected disease which 6Keele University, Keele, UK the risk centres on the threat of financial 7 are core to improving human health, but impoverishment due to catastrophic health O’Neill Institute for National and Global Health Law, neglected in both UHC and GHS.44 Yet, Georgetown University Law Centre, Georgetown expenditures. However, these agendas are University, Washington, District of Columbia, USA many donors expect discrete funding not comprehensively aligned. We recognise 8Harvard University T H Chan School of Public Health, priorities that can be easily measured, divergence between these frameworks; Boston, Massachusetts, USA such as treatment for the big three. between the individual and the collective 9Global Health Policy Unit, University of Edinburgh, Indeed, addressing the potential economic and between domestic and international Edinburgh, United Kingdom threats to national labour force through a priorities. Empirically, we show there are 10Department of Global Public Health and Primary Care, multitude of further health burdens may some overlapping indicators between GHS University of Bergen, Bergen, Norway 11 be one way to align the concept of “risk” and UHC, but there are also a number of University of Sussex, Brighton, UK 12 between UHC and GHS. indicators outside this synergy. To that Centre for Global Public Health, Queen Mary University, London, London, UK A further challenge is resource allocation: extent, the UHC and GHS goals are in ten- 13Georgetown University Edmund A Walsh School of in healthcare systems worldwide, there sion. The research and practice communi- are gaps between available funding and Foreign Service, Washington, District of Columbia, USA ties that represent these two streams need 14Georgetown University O’Neill Institute for National and possible health interventions leading to to engage so that smart strategies can be 3 15 Global Health Law, Washington, District of Columbia, USA priority setting. What are the ethical, 15 identified to improve both aims simultane- University of Edinburgh Division of Health Sciences, political and socioeconomic implications ously using codependent, but distinct pol- Edinburgh, UK of prioritising GHS, which may threaten icy. We suggest HSS as a method to achieve Correspondence to: Dr Clare Wenham HICs, as well as LMICS, rather than both and in doing so build more “equita- ​c.wenham@​ lse.​ ac.​ uk​ addressing Non-Communicable Diseases ble and sustained improvements across (NCDs) relating to the growing tobacco health services and health outcomes”.39 epidemic in Africa or ultraprocessed Yet, we caution that this is not panacea, food in South America? Priority setting but a meaningful step to bringing these This is an Open Access article distributed in accordance implies difficult choices have to be made with the terms of the Creative Commons Attribution global health agendas together in a more (CC BY 4.0) license, which permits others to distribute, and raises important ethical and equity comprehensive mechanism. remix, adapt and build upon this work, for commercial considerations. UHC requires decision use, provided the original work is properly cited. See: makers to agree on criteria and establish Contributors:All authors contributed to a two-day http://creativecommons.org/licenses/by/4.0/. workshop discussion where the content of this paper transparent and fair priority setting was developed. It was subsequently written up by CW, processes.15 Further elaboration is needed RK and DS. All authors reviewed and made comments to understand how concerns for GHS and on the draft text before submission. 1 Ghebreyesus T. World health or ganization, 2017. UHC can be considered within this. Funding:Funding to support the workshop that this Available from: http://www.who.​ int/​ news-​ room/​ ​ Additionally, open definitions of “health paper came out of was provided from the Georgetown commentaries/detail/​ all-​ roads-​ lead-​ to-​ universal-​ ​ systems” and how to measure their strength Global Initiative and the Global Health Governance health-​coverage [Accessed 15 Jun 2018]. Programme at Edinburgh, supported by the Wellcome 2 Ooms G, Beiersmann C, Flores W, et al. Synergies leave the door open for an emphasis on GHS Trust [106635]. and tensions between universal health coverage and entire disease areas (such as NCDs) that and global health security: why we need a second does not address health inequities within a Competing interests:None declared. ‘Maximizing Positive Synergies’ initiative. BMJ Glob Health 2017;2:e000217. country with limited resources. Like Unicef’s Patient consent:Not required. 3 Kutzin J, Sparkes SP. Health systems strengthening, support of selective primary healthcare in the Provenance and peer review:Not commissioned; universal health coverage, health security and 1980s—which was introduced as a means to externally peer reviewed. resilience. Bull World Health Organ 2016;94:2. simplify and actualise primary healthcare 4 Pan American Health Organization (PAHO). Strategy Data sharing statement: No additional data are for universal access to health and universal health goals—and the Gavi and Global Fund coverage. Agenda Item 4.3 CD53/5. Washington D.C: 45 available. approach to HSS, promoting an HSS model PAHO, 53rd Directing Council, 66th Session of the that includes both GHS and UHC means Handling editor :Seye Abimbola Regional Committee of the WHO for the Americs, 2014.

30 Wenham C, et al. BMJ Glob Health 2019;4:e001145. doi:10.1136/bmjgh-2018-001145 UNIVERSAL HEALTH COVERAGE

5 World Health Organization. World health report, 20 Dhillon RS, Kelly JD. Community trust and the ebola 34 UN General Assembly. International covenant on 2010: health systems financing the path to universal endgame. N Engl J Med 2015;373:787–9. economic, social and cultural rights, 1966. coverage. Geneva: WHO, 2010. 21 Whetten K, Leserman J, Whetten R, et al. Exploring 35 Flahault A, Wernli D, Zylberman P, et al. From 6 Musgrove P. Public spending on health care: lack of trust in care providers and the government global health security to global health solidarity, how are different criteria related? Health Policy as a barrier to health service use. Am J Public Health security and sustainability. Bull World Health Organ 1999;47:207–23. 2006;96:716–21. 2016;94:863. 7 Sustainable Development Solutions Network. 22 Yamey G. Can Japan rouse the g7 nations to action 36 Davies SE. Securitizing infectious disease. Int Aff Indicators and a monitoring framework: launching on universal health coverage?, 2015 Available from: 2008;84:295–313. a data revolution for the sustainable development https://blogs.​ bmj.​ com/​ bmj/​ 2015/​ 12/​ 22/​ gavin-​ ​ 37 World Health Organization. The Tallinn Charter: goals, 2018. Available from: http://​indicators.​report/​ yamey-can-​ japan-​ rouse-​ the-​ g7-​ nations-​ to-​ action-​ on-​ ​ health systems for health and wealth. Copenhagen: targets/​3-​8/ [Accessed 17Jun 2018]. universal-​health-​coverage/ [Accessed 31 May 2018]. WHO Regional Office for Europe, 2008. 8 World Health Organization World Health Report 23 Heymann DL, Chen L, Takemi K, et al. Global 38 Kieny MP, Bekedam H, Dovlo D, et al. Strengthening 2007. A safer future: global public health security in health security: the wider lessons from the west health systems for universal health coverage and 21st century. Geneva, 2007. African Ebola virus disease epidemic. Lancet sustainable development. Bull World Health Organ 9 Kamradt-Scott A, Wroczynzki M. Global health 2015;385:1884–901. 2017;95:537–9. security and the BMJ. London: BMJ Global Health 24 World Health Organization. Tracking universal health 39 World Health Organization. Everybody’s business- Security, 2017. coverage: 2017 Global Monitoring Report. Geneva, strengthening health systems to improve health 10 Buzan B. People, states & fear: an agenda for 2017. outcomes: WHO’s framework for action. Geneva, international security studies in the post-cold war 25 Saywell T, Fowler G, Crispin S. The Cost of SARS: 2007. era. ECPR Press, 2008. $11Bn and Rising. Dow Jones Far Eastern Economic 40 Farmer P, 2016. Solving the ebola outbreak: paul 11 Aldis W. Health security as a public health Review, 2003. farmer and the four s’s. Talk given at university of concept: a critical analysis. Health Policy Plan 26 Rassy D, Smith RD. The economic impact of H1N1 Notre Dame. Available from: https://​science.​nd.​edu/​ 2008;23:369–75. on Mexico’s tourist and pork sectors. Health Econ news/solving-​ the-​ ebola-​ outbreak-​ paul-​ farmer-​ and-​ ​ 12 Rushton S. Global health security: security for whom? 2013;22:824–34. the-​four-​ss/ [Accessed 1Jul 2018]. Security from what? Polit Stud 2011;59:779–96. 27 Evans D. The economic impact of the 2014 Ebola 41 Kluge H, Martín-Moreno JM, Emiroglu N, et al. 13 World Health Organization. International health Epidemic: short-and medium-term estimates for West Strengthening global health security by embedding regulations (2005). Geneva, 2005. Africa. Washington DC: World Bank Group, 2014. the Regulations requirements into national health 14 Office for the High Commissioner on Human Rights. 28 National Academies of Medicine. Global health systems. BMJ Glob Health 2018;3.e000656. CESCR general comment no. 14: the right to the highest risk framework: the neglected dimension of global 42 World Health Organization, 2018. Fact sheet on attainable standard of health (Art 12. Geneva, 2000. security: a framework to counter infectious disease universal health coverage. Available from: http://www.​ 15 World Health Organization. Making fair choices on crises. Washington DC, 2016. who.int/​ en/​ news-​ room/​ fact-​ sheets/​ detail/​ universal-​ ​ the path to universal health coverage: final report of 29 Smith J. Global health security: a flawed SDG health-coverage-(​ uhc)​ [Accessed 1 Jul 2018]. the WHO consultative group on equity and universal framework. Lancet 2015;385:2249. 43 Hogan D, Reza Hosseinpoor A, Boerma T. Developing health coverage. Geneva, 2014. 30 Chan M. WHO director-general’s speech to the regional an index for the coverage of essential health services. 16 Global Health Security Agend. Global heath security committee for the western pacific. Manila, 2014. Geneva: World Health Organization, Department of agenda, 2018. Available from: https://www.​ 31 Dunn JT, Lesyna K, Zaret A. The role of human rights Evidence, Information and Research, 2016. ghsagenda.​org/ [Accessed 8 Oct 2018]. litigation in improving access to reproductive health 44 Schmidt H, Gostin LO, Emanuel EJ. Public health, 17 Jain V, Alam A. Redefining universal health coverage care and achieving reductions in maternal mortality. universal health coverage, and sustainable in the age of global health security. BMJ Glob Health BMC Pregnancy Childbirth 2017;17(Suppl 2):367. development goals: can they coexist? Lancet 2017;2:e000255. 32 General Assembly UN. International Covenant on 2015;386:928–30. 18 Ezzati M, Lopez AD, Rodgers A, et al. Selected major Economic, Social and Cultural Rights: CESCR (2000) 45 Storeng KT. The GAVI Alliance and the ‘Gates risk factors and global and regional burden of General Comment 14: The right to the highest approach’ to health system strengthening. Glob disease. Lancet 2002;360:1347–60. attainable standard of health; CESCR (1990). Gen Public Health 2014;9:865–79. 19 Heymann DL, Chen L, Takemi K, et al. Global health 1966. December 16. security: the wider lessons from the west African Ebola 33 UN General Assembly. International covenant on civil Cite this as: BMJ Glob Health 2019;4:e001145 virus disease epidemic. Lancet 2015;385:884–1901. and political rights, 1966. http://dx.doi.org/10.1136/bmjgh-2018-001145

Wenham C, et al. BMJ Glob Health 2019;4:e001145. doi:10.1136/bmjgh-2018-001145 31 UNIVERSAL HEALTH COVERAGE

How moves towards universal health coverage could encourage poor quality drugs Universal health coverage depends on affordable medicines. But pushing down prices without also investing in quality assurance will increase the sale of substandard and falsified drugs, warns Elizabeth Pisani

any governments in middle ingredients or packaging, or skipping some compromising quality, and many aren’t income countries are quality assurance steps. prepared to do that. But most are not prepared working hard to deliver on The result can be drugs that are so sloppily to disappoint shareholders by reducing profit political promises that all made that they don’t dissolve properly in margins either. their citizens will have access the body; that degrade before the patient Procurement officials in the health ministry Mto quality health services, without being takes them, sometimes because of cheap but might decide that costs plus a 10% profit impoverished. They are finding that universal inappropriate packaging or handling; or that represents a “fair” price, but many drug health coverage (UHC) doesn’t come cheap. are dangerously lacking in active ingredients. producers and distributors of innovator Indonesia’s national health insurance Similar problems have been reported from and generic drugs calculate globally. They scheme, for example, has given out 223.4 India, one of the biggest producers of cheap compare that margin with what they can million health cards since its inception in drugs.8 make on other products or in other markets. 2014. Nationwide, 73% of households said at Drug regulators are supposed to spot this If procurement practices drive prices down least one household member had some health kind of corner cutting before procurement in one market, multinational companies just insurance in 2018, up from 52% in 2013.1 2 agencies buy them—and they do, but only if pull (or stay) out of the country. Domestic Yet the scheme has been in permanent deficit; they have the right people, money, technology, producers, who often have higher costs, by 2018 it had a shortfall of 23 trillion rupiah skills, and incentives. Our study found that sometimes stop making comparatively (£1.3bn; €1.5bn; $1.6bn.3 Turkey’s drug regulator, which employs underpriced products entirely. Such deficits lead to belt tightening. over 3000 well trained inspectors, is widely The Romanian government, struggling Globally, about a quarter of all health considered to provide effective oversight.7 with large deficits in its national health spending is on drugs. In poorer countries the But in the mad dash to reach UHC on a insurance programme, in 2009 introduced proportion is higher, and patients typically shoestring, many middle income countries a system designed to cut drug prices to the foot more of the bill.4 As governments move underinvest in developing regulatory capacity. European minimum or below. After a radical towards UHC, they increasingly pay for drugs Other forces are at play, too. Domestic drug revision of the price caps in 2015, prices fell that used to be paid for by patients—and look makers are sometimes protected by local by 16-25% overnight, while European Union for ways to push prices down. authorities, who have made promises to voters rules allowed drugs purchased in Romania Cheaper drugs should mean more people about jobs. Regulators in China and Indonesia to be resold in other countries. As a result, effectively treated for the same budget, taking told us that politicians, unwilling to sacrifice manufacturers withdrew about 2000 of 6200 countries towards UHC. There’s plenty of votes or tax income from industry, have authorised medicines from the Romanian room for belt tightening. Generic and branded discouraged thorough inspection of factories market. Patients in Romania now look for drug makers often charge whatever they can.5 or warehouses. those drugs on the internet, and doctors have Inefficient procurement and plain old corruption And there’s no effective way to ensure that reported using smuggled products.7 push prices up; some poorer countries pay 30 imported products are well made. The global Vacuums left by product withdrawal are times more than the international reference price drugs market operates on a “buyer beware” easily filled by repackaged expired drugs or for basic generic drugs.6 system—national medicines regulators stolen or smuggled products that avoid taxes, But countries with under-resourced don’t have to assure the quality of products import duties, or registration costs—and health budgets seeking to push down for export. Most imported drugs are waived bypass regulatory oversight. The gap may also prices should be careful what they wish through with paper based assurances; well be filled by fakes, which cost almost nothing for. Recent research in China, Indonesia, resourced regulators only test a tiny fraction to make. WHO reports fakes in all classes of Romania, and Turkey found evidence of imported drugs. medicines, worldwide.11 that drug manufacturers and distributors Around 30% of countries globally, according High prices can also effectively create react quickly to keep profits as high as to the World Health Organization’s latest shortages for uninsured patients if they can’t possible,7 potentially leaving patients count, don’t have the capacity to regulate drugs afford to buy a drug. In theory, UHC should exposed to substandard drugs, and creating properly, even at home.9 They take what they reduce this risk, by covering the cost of more opportunities for criminals to sell fake drugs. are sent, and it is not always first rate. A 2017 drugs. But even well resourced health systems WHO summary of studies going back a decade restrict access to drugs people need or want. The downside of cheap drugs estimated that one in 10 anti-infective drugs In England, for example, the NHS does not One way to maintain profits is to cut sold in low and middle income countries didn’t cover pre-exposure prophylaxis for HIV or production costs—for example, by shifting meet minimum quality standards.10 the latest treatments for cystic fibrosis. In manufacturing to cheaper locations or poorer countries, health insurance cards increasing worker productivity. Some Filling the vacuum may facilitate access to health services but manufacturers also mentioned more worrying The most efficient producers probably can’t leave patients to pay for the most expensive measures, such as switching to cheaper cut production costs much further without prescriptions.

32 doi: 10.1136/bmj.l5327 | BMJ 2019;366:l5327 | the bmj UNIVERSAL HEALTH COVERAGEFEATURE

“Affordability shortages” help drug falsifiers (often branded) drugs. Limited field surveys Convention requires countries to accept flights in two ways. Firstly, falsifiers seek to maximise suggest substandard and degraded drugs are from other nations only if their regulatory and profits, and higher prices mean more profits more often found among lower cost generics, safety procedures meet standards set by the for similar outlay and risk. Secondly, falsifiers especially where regulation is weak.15 International Civil Aviation Organisation. are criminals, and patients who can’t afford Generics are essential to every national It also allows for cross national inspection, the prices at regulated outlets chase bargains health system, and most work just fine, even quality assurance, incident reporting, and on the internet, in street markets, or in buyers’ in less regulated markets. No one wants to investigation.17 There’s no sign that national clubs, which are harder to police. risk undermining public confidence in the sovereignty has suffered as a result. pharmaceutical bedrock on which UHC must Thirdly, countries should adjust health Selling the idea of quality be built. But unless quality assurance is strong budgets in response to other policy choices. Changes in health financing in efforts to enough, patients and doctors, who have no For example, few nations can produce achieve UHC in middle income countries way of knowing what’s actually in a pill or quality assured drugs as cheaply as the great may also eat into the revenues and profits of a vial, will continue to rely on sometimes exporting powerhouses, so if a country decides health providers serving insured patients. Less unreliable signals of quality such as price and to promote local jobs by buying medicines scrupulous providers may choose to top up brand. from domestic producers, it will have to pay their earnings by encouraging patients to buy This creates a headache for governments more. Similarly, stricter environmental rules drugs “off plan,” often in the name of quality. striving to achieve UHC. Delivering better for factories will push up production costs, This can increase patients’ risk of getting quality assurance requires investment in and thus prices. If countries want to pursue falsified products. technology and skills, which countries need such policies and deliver UHC, they’ll have to An example comes from Indonesia, where to build into budgets and practice. It’s an increase budget allocations for drugs. in 2016 over 1000 children received fake inescapable part of achieving UHC. Taking account of the price of quality in vaccines, supposedly made by multinational policy and procurement decisions will not producers GSK and Sanofi. Domestically Access to drugs that work increase access to drugs, but it could help to produced, WHO prequalified vaccines were The market for quality assured drugs is universally available free but earned doctors disrupted by a mismatch between a largely ensure that the drugs that are on the market only a tiny fee. Paediatricians at some private socialised demand side, and an almost actually work as intended. hospitals preferred instead to offer imported entirely capitalist supply side. In countries Elizabeth Pisani is an epidemiologist who researches the interaction between politics, human behaviour, vaccines at up to $40 a shot, exploiting public aiming for UHC, most decisions about buying and health. She currently focuses on the political tendency to consider high cost an indication of drugs are made by governments aiming to and economic drivers of substandard and falsified quality. To maximise earnings, some doctors minimise spending while protecting local jobs medicines, with a particular interest in countries that are bought the vaccines at cut price from roving and revenues. These goals are often internally rapidly scaling up access to health services. This essay is informed by research in four countries, supported by salespeople. In fact, these fake vaccines were incompatible and certainly don’t align with the Wellcome Trust and Erasmus School of Health Policy made in a garage in a Jakarta suburb.12 the aims of drug companies: to maximise and Management. global profits. Until that changes, substandard This article is part of a series commissioned by The What’s the damage? and falsified drugs will continue to flourish. BMJ based on an idea from the Harvard Global Health Most fake drugs contain little or no active Existing efforts to change incentives for drug Institute. The BMJ retained full editorial control over ingredient, and many contain toxic external peer review, editing, and publication. Harvard discovery, production, and procurement may Global Health Institute paid the open access fees. substances. These will obviously fail to cure improve access to affordable, quality assured and may harm. It’s harder to assess the damage drugs, but will take time.16 Three interim Competing interests: I have read and understood BMJ done by substandard drugs. Occasionally, measures might help protect drug quality in policy on declaration of interests and have no relevant interests to declare. drugs will be so badly made that they will the meantime. poison and kill people, but that’s mercifully Firstly, national discussions about fair Provenance and peer review: Commissioned; rare.13 Generally, substandard drugs simply prices should consider the globalised market. externally peer reviewed. fail to prevent, treat, or cure disease as they Recent efforts to promote open pricing could This work draws heavily on the research and thinking of Adina-Loredana Nistor, Amalia Hasnida, Koray should. Money is wasted, and patients are narrow the price difference between markets, Parmaksiz, Jingying Xu, Maarten O Kok, and Pernette weakened and may die unnecessarily, though especially if countries also share medicine Bourdillon-Esteve. I thank Ibrahim Abubakar, Martin death is usually blamed on underlying illness registration and procurement mechanisms. McKee, Gillian Buckley, and Paul Simpson for helpful and the role of poor quality drugs often goes But expect resistance from politicians or comments on an earlier draft. unrecognised. bureaucrats (claiming national sovereignty) Elizabeth Pisani, associate professor, epidemiology If the disease is infectious, prevention who see in UHC an opportunity for kickbacks Erasmus School of Health Policy and Management, Rotterdam, The Netherlands, and King’s College failure and longer illness can mean more from national procurement contracts. London, UK transmission. Subtherapeutic doses of Secondly, increase regulatory resources Correspondence to: [email protected] anti-infectives contributes to antimicrobial in countries that export to low and middle resistance, which can spread globally, thus income markets. This will also ruffle feathers, reducing the effectiveness of quality assured including among regulators in importing drugs in well regulated markets.14 countries who worry about sovereignty and This is an Open Access article distributed in accordance These concerns don’t seem to ignite high lost revenue. There are successful precedents, with the Creative Commons Attribution Non Commercial profile, activist led campaigns. We know however. WHO’s prequalification programme (CC BY-NC 4.0) license, which permits others to neither the prevalence nor the distribution of has increased quality assurance at source distribute, remix, adapt, build upon this work non- poor quality drugs. The logic of the market, for drugs to treat HIV, tuberculosis, malaria, commercially, and license their derivative works on different terms, provided the original work is properly and information from regulators, suggests and reproductive health, for example. In cited and the use is non-commercial. See: http:// that falsifiers more often target high value the aviation industry, the 1944 Chicago creativecommons.org/licenses/by-nc/4.0/. the bmj | BMJ 2019;366:l5327 | doi: 10.1136/bmj.l5327 33 UNIVERSAL HEALTH COVERAGE

7 Pisani E, Nistor A-L, Hasnida A, Parmaksiz K, Xu J, b876f34f36964122bcfe0408002d0415/vaccine- Kok MO. Identifying market risk for substandard scandal-highlights-indonesian-health-system-woes. and falsified medicines: an analytic framework 13 Lahore High Court. The pathology of negligence: 1 StatisticsIndonesia. Survei Sosial Ekonomi Nasional based on qualitative research in China, report of the Judicial Inquiry Tribunal to determine the 2018. 2019. Indonesia, Turkey and Romania. Wellcome causes of deaths of patients of the Punjab Institute of 2 Kesehatan BPJS. Jumlah Peserta program JKN. 2019. Open Res 2019;4:70. doi:10.12688/ Cardiology, Lahore in 2011-2012. 2012. http://apps. https://bpjs-kesehatan.go.id/bpjs/jumlahPeserta. wellcomeopenres.15236.1 who.int/medicinedocs/en/m/abstract/Js22131en. 3 Kesehatan BPJS. Laporan Pengelolaan program Tahun 8 Eban K. Bottle of lies: the inside story of the generic 14 Pisani E. Antimicrobial resistance: what does medicine 2017 dan Laporan Keuangan Tahun 2017 (Auditan). drug boom. Ecco Press, 2019. quality have to do with it? Antimicrobial Review, 2018. https://bpjs-kesehatan.go.id/bpjs/index.php/ 9 World Health Organization. Improving the quality of 2015. http://amr-review.org/sites/default/files/ arsip/categories/Mzg/laporan-keuangan medical products for universal access. WHO Essential ElizabethPisaniMedicinesQualitypaper.pdf. 4 Silverman R, Keller JM, Glassman A, Chalkidou Medicines and Health Products. www.who.int/ 15 Bate R, Jin GZ, Mathur A. Falsified or substandard? K. Tackling the triple transition in global health medicines/regulation/fact-figures-qual-med/en. assessing price and non-price signals of drug procurement. Center for Global Development. 2019. 10 World Health Organization. A study on the public quality. J Econ Manage Strategy 2015;24:687-711. www.cgdev.org/sites/default/files/better-health- health and socioeconomic impact of substandard doi:10.1111/jems.12114 procurement-tackling-triple-transition.pdf. and falsified medical products. 2017. http://who. 16 Wirtz VJ, Hogerzeil HV, Gray AL, et al. Essential medicines 5 Hill AM, Barber MJ, Gotham D. Estimated costs of int/medicines/regulation/ssffc/publications/Layout- for universal health coverage. Lancet 2017;389:403- production and potential prices for the WHO Essential SEstudy-WEB.pdf. 76. doi:10.1016/S0140-6736(16)31599-9 Medicines List. BMJ Glob Health 2018;3:e000571. 11 World Health Organization. WHO Global Surveillance 17 Havel BF, Sanchez GS. The international law regime doi:10.1136/bmjgh-2017-000571 and Monitoring System for substandard and falsified for aviation safety and security. In: The principles 6 AfRx Consulting. Background research and landscaping medical products. 2017. www.who.int/medicines/ and practice of international aviation law. Cambridge analysis on global health commodity procurement. regulation/ssffc/publications/GSMS_Report.pdf. University Press, 2014. Center for Global Development, 2018. www.cgdev.org/ 12 Karmini N, Mason M. Vaccine scandal sites/default/files/CGD-procurement-background-afrx- highlights Indonesian health system woes. Cite this as: BMJ 2019;366:l5327 research-landscape.pdf. Associated Press. 2016. https://apnews.com/ http://dx.doi.org/10.1136/bmj.l5327

34 doi: 10.1136/bmj.l5327 | BMJ 2019;366:l5327 | the bmj UNIVERSAL HEALTH COVERAGE

Universal quality health coverage—committing to a healthier and more productive society High quality healthcare can be achieved in all settings with good leadership, robust planning, and intelligent investment, say Jeremy Veillard and colleagues

igh quality healthcare involves Improving access to care, especially rich foundation from which to rapidly scale the right care, at the right for poor people, through universal health up a quality revolution. For the first time, time, in the right place, and coverage is not enough to achieve better the report reviews evidence available for by the right care provider, health outcomes. This is the simple 23 distinct interventions that governments, while minimising harm and and powerful reminder from a new managers, and clinicians can use to Hresource waste and leaving no one behind. report coauthored by the World Health improve quality of care. Among those, Unfortunately, many countries around the Organization the World Bank Group, and seven categories of interventions stand world fail to meet these requirements and the Organisation for Economic Cooperation as priorities: changing clinical practice at struggle to provide healthcare services and Development.5 It is the first global the frontline; setting standards; engaging that deliver clinical value to patients, are report coauthored by the three multilateral and empowering patients, families, and safe, and meet the needs and preferences organisations. communities; information and education of patients. The report calls for urgent action from for healthcare workers, managers, and Poor quality healthcare prevails governments, clinicians, patients, civil policy makers; use of continuous quality in countries at all levels of economic society, and the private sector to help improvement programmes and methods; development, manifesting every day in rapidly scale up quality healthcare services establishing performance based incentives inaccurate diagnoses, medication errors, for universal health coverage. To start with, (financial and non-financial); and inappropriate or unnecessary treatment, governments should develop national legislation and regulation. and inadequate or unsafe clinical quality policies and strategies that address Each country requires different sets of facilities or practices. The implications are the foundations of quality health systems. interventions to improve quality of care— devastating for patients and their families. Building quality health services requires a depending on its quality baseline, resources In low and middle income countries, for culture of transparency, engagement, and available, capacities and capabilities, example, 10% of hospitalised patients can openness about results, which should be and needs and expectations from the expect to acquire an infection during their promoted in all societies. National quality populations served. The report describes stay, compared with 7% in high income policies and strategies should ensure that how four countries with vastly different countries.1 The World Bank’s service healthcare workers are motivated and contexts—Canada, Ethiopia, Mexico, and delivery indicators found that healthcare supported to provide quality care; that Sudan—are doing so systematically. workers in seven low and middle income healthcare services are accessible and well Of course, quality care requires some African countries were able to make equipped; that action is taken to ensure investment, but it is affordable, especially accurate diagnoses only one third to three that medicines, devices, and technologies when the costs and consequences of quarters of the time, and clinical guidelines are safe in design and use; that information poor quality are considered. Many of the for common conditions were followed less systems continuously measure, monitor, interventions to improve quality—think of than 45% of the time on average.2 A recent report, and drive better quality care; and checklists or basic hygiene, for example— study found that, despite access to care finally, that the way healthcare providers are inexpensive and within reach for all improving around the world, the clinical are paid for encourages and enables quality countries. The returns are plentiful— benefits for patients and populations of care. better individual and population health, remain limited: in eight high mortality High quality healthcare for all might more productive workers, and pupils countries in Africa and the Caribbean, seem ambitious, but it can be achieved in that perform better in school and will effective coverage averaged 28% for all settings with good leadership, robust contribute better to the economy. In other antenatal care, 26% for family planning, planning, and intelligent investment. In words, investment in quality healthcare and 21% for paediatric care.3 Uganda, for example, a model involving contributes to growth in human capital Poor quality healthcare also has major citizens and communities in the design of and economic development. So striving costs for people’s lives, health systems, healthcare services has improved a range for universal quality health coverage is not and societies. In high income countries of indicators, including a 33% reduction just an investment in better health—it is a alone, harmful medical errors and in child mortality.6 Costa Rica has also commitment to building a healthier and preventable complications account for achieved remarkable improvements in more productive society. 15% of hospital costs.4 Globally, misuse primary care quality through a carefully Jeremy Veillard is a programme manager in global and inappropriate use of antimicrobials planned, implemented, and resourced engagement at the World Bank’s health, nutrition, is fuelling the rise of antimicrobial improvement strategy focused on and population global practice, where he focuses on resistance, leading to considerable quality.7 service delivery redesign and quality of care. He is also an assistant professor at the Institute of Health additional health spending, affecting Around the world, lessons abound on Policy, Management, and Evaluation at the University labour supply and productivity. what works and what does not, providing a of Toronto, Canada. the bmj | BMJ 2019;366:l5489 | doi: 10.1136/bmj.l5489 35 UNIVERSAL HEALTH COVERAGE

Edward Kelley is director of the department of service Niek Klazinga is senior health economist/policy in eight high-mortality countries. BMJ Glob delivery and safety at the World Health Organization. analyst at the OECD, working on the measurement and Health 2017;2:e000424. doi:10.1136/ In this role, he leads WHO’s efforts to strengthen the improvement of healthcare quality and outcomes. He bmjgh-2017-000424 safety, quality, integration, and people centredness is also professor of social medicine at the Amsterdam 4 Organisation for Economic Cooperation and of health services globally and is the lead for WHO’s University Medical Centre, The Netherlands. Development. The economics of patient safety: work on strengthening health systems and security. strengthening a value-based approach to reducing The opinions expressed and arguments employed patient harm at national level. 2017 https://www. Sepideh Bagheri Nejad is a technical officer herein are solely those of the author and do not oecd-ilibrary.org/social-issues-migration-health/the- responsible for knowledge management at the necessarily reflect the official views of the OECD or of economics-of-patient-safety_5a9858cd-en department of service delivery and safety at the its member countries. 5 World Health Organization, Organisation for World Health Organization. She coordinated the This article is part of a series commissioned by The Economic Cooperation and Development, World development of the joint WHO-OECD-World Bank BMJ based on an idea from the Harvard Global Health Bank. Delivering quality health services. A global report Delivering quality health services: a global Institute. The BMJ retained full editorial control over imperative for universal health coverage. 2018. imperative for universal health coverage. editing and publication. https://www.who.int/servicedeliverysafety/quality- report/en/ Francesca Colombo is head of the OECD health Competing interests: None declared. 6 World Vision International. Citizen voice and action: division. OECD work on health provides internationally civic demand for better health and education comparable data on health systems and economic Correspondence to: J Veillard services. 2012. https://www.escr-net.org/ analysis of health policies, advising policy makers, [email protected] node/366893 stakeholders, and citizens on how to respond to 7 Organisation for Economic Cooperation and demands for more and better care. 1 Organisation for Economic Cooperation and Development. Health at a glance 2017: OECD Development. OECD reviews of health systems: Tim Evans is the senior director for health, nutrition, indicators. 2017. http://www.oecd.org/health/ Costa Rica 2017. 2017. http://www.oecd.org/ and population at the World Bank Group and has health-systems/health-at-a-glance-19991312.htm countries/costarica/oecd-reviews-of-health- been active in the international health arena for more 2 World Bank. DataBank: service delivery indicators. systems-costa-rica-2017-9789264281653-en. than 20 years. He earned his DPhil in agricultural https://databank.worldbank.org/data/source/ htm. economics at Oxford and pursued medical and service-delivery-indicators postgraduate studies at McMaster and Harvard 3 Leslie HH, Malata A, Ndiaye Y, Kruk ME. Cite this as: BMJ 2019;366:l5489 Universities. Effective coverage of primary care services http://dx.doi.org/10.1136/bmj.l5489

36 doi: 10.1136/bmj.l5489 | BMJ 2019;366:l5489 | the bmj UNIVERSAL HEALTH COVERAGE

Non-communicable diseases must be part of universal health coverage As the world’s leading cause of preventable morbidity and disability, non-communicable diseases are central to achieving universal health coverage, says Sania Nishtar

he move towards universal health drug lists and national drug procurement Technological innovations ranging coverage (UHC) has been spurred systems. Access to disease preventing from telemedicine and mobile health to on by the risk of a pandemic and vaccines is a key component of UHC. The artificial intelligence and the internet of the premise of the sustainable human papillomavirus (HPV) vaccine, medical things, as well as personalisation development agenda. But, as coun- for example, is a cost effective measure and on-demand healthcare, are largely Ttries move towards UHC, their frame of refer- to prevent cervical cancer, a non- drawn on non-communicable diseases. ence often converges around services focused communicable disease that needlessly As health systems of the future are shaped on maternal and child health and infectious kills over 270 000 women each year. with these changes, and UHC plans draw on diseases. This is likely to be because of the HPV vaccination is being rolled out them to make health systems more efficient programmes that were established in the era from Australia to Rwanda. Coupled with and effective, the integration of non- of the millennium development goals. increased screening, HPV elimination is now communicable disease care will become Non-communicable diseases must be feasible and, with support from countries more feasible. But there must be political part of UHC frameworks—they are the and organisations like Gavi, the Vaccine will to do so. world’s biggest killers and the leading cause Alliance, the vaccine can be accessible Integration with UHC is critical of preventable morbidity and disability. across the globe. Similarly, the hepatitis B for closing the services gap in non- This does not mean a vertical solution. virus (HBV) is the cause of 43% of deaths communicable diseases and tackling the Opportunities exist to negotiate policy from liver cancer globally. Liver cancer rates rates of unnecessary deaths, disability, and space for non-communicable diseases in tend to be highest in countries with minimal illness. Although essential, however, it is each health system’s six “building blocks,” vaccination programmes. If administered not enough. Non-communicable diseases and therefore in UHC plans. early, the HBV vaccine can save lives. are uniquely responsive to changes in In terms of service delivery and financing, Complexities in the diagnosis of non- the environments in which people live. where free at the point of delivery measures communicable diseases necessitate Factors that contribute heavily to non- are being adopted to achieve UHC, non- surveillance of risk factors rather than communicable disease risk factors—such as communicable disease services must be diseases. This can be integrated with unhealthy diet, physical inactivity, tobacco included in UHC entitlements. Where existing population based surveys such and alcohol use, and air pollution—are insurance schemes are being used to as the Living Standards Measurement beyond the purview of the health system, finance UHC, non-communicable diseases Survey, Demographic Health Survey, or hence the urgent need for multisectoral must be included in the public benefits the Multiple Indicator Cluster Survey. action. package. And where social protection These are routinely conducted in over 100 Prevention strategies, such as the WHO programmes are the instrument of UHC, countries. Integrating surveillance of non- recommended “Best Buys,”4 and more they must acknowledge that much of the communicable disease with these surveys recent initiatives, such as WHO’s move risk of catastrophic health expenditure is would remove the need for standalone to remove trans fats from the global food derived from non-communicable diseases. surveys. This could be supplemented with chain,5 are critically important in tackling Primary healthcare must be reoriented cancer registries and non-communicable non-communicable diseases. They could towards chronic care. This can be done disease sensitive metrics in service help us save 8 million lives by 2030.6 readily in settings where HIV chronic management information systems. To be successful, governments must care platforms have been established, as The inclusion of training modules unbundle the term “multisectoral.” More these provide an opportunity to jumpstart on non-communicable diseases into broadly, an overarching change is needed nascent non-communicable disease the curriculums of ongoing education so that governments choose to prioritise programmes. A recent partnership between for health workers can help prime long term sustainability over short term the US President’s Emergency Plan for AIDS professionals in the new realities in public gratification and to calculate the true cost Relief and AstraZeneca,1 aimed at offering health. Community health workers and borne by societies in the future instead HIV/AIDS and hypertension services in an nurses have critical roles in giving lifestyle of just the price of actions and policies integrated manner, is a step in the right advice and support in reducing stigma today. direction. Recent initiatives, such as the and discrimination, which is especially The global community has a respon­ Defeat NCD Partnership and Resolve,2 important when it comes to mental health. sibility to facilitate this transformation.­ can help accelerate the integration of Partnerships with the nursing community Countries have become accustomed to non-communicable diseases into primary can be rewarding. The Nursing Now project the politically attractive clear “asks,” healthcare. is looking at ways to further leverage their fiscal envelopes, and hand holding of Drugs for non-communicable diseases strengths, including collaboration with the the millennium development goal era. should be included in national essential non-communicable disease community.3 There should be no escaping a strong the bmj | BMJ 2019;366:l5401 | doi: 10.1136/bmj.l5401 37 UNIVERSAL HEALTH COVERAGE

global push to tackle non-communicable Sania Nishtar is co-chair of the WHO independent 3 World Health Organization. Nursing Now Campaign. diseases. A strong institutional base, a high level commission on non-communicable https://www.who.int/hrh/news/2018/nursing_now_ diseases. campaign/en/ clear strategy, and adequate funding are 4 World Health Organization. “Best buys” and other critically needed. Competing interests: None. recommended interventions for the prevention and Non-communicable diseases will be This article is part of a series commissioned by control of noncommunicable diseases. https://apps. The BMJ based on an idea from the Harvard Global who.int/iris/bitstream/handle/10665/259232/ profiled at the United Nations General HealthInstitute. The BMJ retained full editorial control WHO-NMH-NVI-17.9-eng.pdf;jsessionid=B4EC0B35 Assembly high level meeting in September. over editing and publication. 0648507179248CB99EAEA3AD?sequence=1 Never has there been a wider chasm in Correspondence to: [email protected] 5 World Health Organization. WHO plan to eliminate @SaniaNishtar industrially-produced trans-fatty acids from global food public health between our knowledge supply. 14 May 2018. https://www.who.int/news-room/ detail/14-05-2018-who-plan-to-eliminate-industrially- about the burden of disease coupled with 1 PEPFAR and AstraZeneca launch partnership across produced-trans-fatty-acids-from-global-food-supply evidence of what works on one hand, HIV and hypertension services in Africa. 8 Sep 2016. 6 World Health Organization. Replace. Trans fat-free by https://www.astrazeneca.com/media-centre/press- and inaction on the other. We must act 2030. https://www.who.int/nutrition/topics/replace- releases/2016/pepfar-and-astrazeneca-launch- decisively to bridge that. The future will transfat partnership-across-hiv-and-hypertension-services-in- hold us accountable for our actions. Words africa-080920161.html Cite this as: BMJ 2019;366:l5401 are not enough. 2 Defeat-NCD Partnership. https://defeat-ncd.org/ http://dx.doi.org/10.1136/bmj.l5401

38 doi: 10.1136/bmj.l5401 | BMJ 2019;366:l5401 | the bmj UNIVERSAL HEALTH COVERAGE

Building more effective health coverage in Argentina Argentina needs to provide actual rather than aspirational universal health coverage, says Adolfo Rubinstein

rgentina, an upper middle benefit plan, which is guaranteed for all provinces.6 The maternal death rate has a income country, has a devel- beneficiaries of social or private insurances, national average of 3.4 per 10 000 of the oped healthcare system that and a reimbursement fund for costly population and an eightfold difference performs well compared with healthcare technologies. But healthcare between poor and rich provinces. Once other countries in Latin Amer- coverage in Argentina is unequal: more we have reached a consensus on which Aica. Health outcomes lag behind the coun- than 60% of the poorest 20% of the conditions should be included as priority try’s potential, however, considering it is population have no insurance, compared health problems, we can set clinical care one of the leaders in the region in terms of with less than 10% in the wealthiest 20% pathways with indicators and quality health spending per capita, which is 10% of of the population.4 targets across all sectors. its gross domestic product.1 We are now working on a strategy to Finally, we are building a primary care The United Nations has set universal advance the integration of healthcare oriented healthcare system. Argentina has health coverage (UHC) as a target for coverage among subsectors. This reform a hospital centred model, which is poorly sustainable development goal 3.8, to be faces many challenges to achieve effective focused on primary care. There is strong achieved by 2030.2 The dimensions of UHC UHC and to accomplish the outcome evidence to support the positive effects of run along three axes: the population that is improvements that the UHC goals set out primary healthcare on improving health covered by pooled funds; the proportion of to achieve.5 We are focusing on four key quality and outcomes.7 Our plan, inspired direct health costs covered by pooled funds; issues. by Brazil’s family health strategy, is based and the health services that are included.3 Firstly, we are setting up provincial on identifying the population served in Although everyone in Argentina is public insurance schemes by creating each primary care catchment area and entitled to receive healthcare, free of subsidised plans for uninsured people at assigning a defined population to core charge, in public facilities, UHC is nominal a provincial level. We are relying on one family health teams.8 The process will be and implicit rather than actual and explicit, of the national programmes that offers leveraged through specific financial and which does not necessarily translate into a good platform from which to launch non-financial incentives for population effective or quality care. In reality, two this strategy. The SUMAR programme, empanelment, achieving quality targets, thirds of Argentina’s population of 45 sponsored by the World Bank, has made and enforcing adequate referrals through million also have social health or private huge progress in advancing UHC strategy local and regional integrated networks of insurance. This leaves another third (an in the public sector, by strengthening the healthcare. estimated 16.5 million people) with no public insurance schemes in a traditional Although Argentina has achieved explicit coverage. Therefore, the public supply driven public healthcare sector. nominal UHC, it still needs to work on sector, funded by taxes, serves as a sort of Secondly, we are creating a transparent achieving effective UHC. This milestone is re-insurance for the whole population. process for setting priorities for a health one of the current government’s national Many of the shortcomings of Argentina’s benefit plan between the national level and priorities. Our ultimate goal is to provide healthcare system come from its segmented the provinces. These priorities will be based actual rather than aspirational UHC, and highly fragmented system. The social on clinical effectiveness, cost effectiveness, improving not only health outcomes, but health insurance sector, which is dominant feasibility, budget impact, opportunity, and also its distribution among different groups, in Argentina, is composed of around 300 social preferences. The package should thereby ensuring better quality healthcare different funds (“obras sociales”). These eventually converge towards the health and equity for every Argentinian. vary in scope and size and are mostly benefit plan of social security. In this Adolfo Rubinstein is the minister of health of managed by trade unions, white collar regard, the imminent launch of a federal Argentina. He is a family doctor, professor of public personnel, and civil servants at a federal agency for health technology assessment health, and professor of family medicine at the school level and in the provinces. Five million will help set objectives and transparent and of medicine of the University of Buenos Aires. His research was focused on epidemiology and preven- older people are covered by a social explicit criteria to define the health benefit tion of non-communicable diseases, implementation health insurance fund for retired workers. plan across the different health sectors. science, health services and policy research, and Overall, social security provides healthcare Thirdly, we aim to reduce disparities in economic evaluations of interventions of non-commu- coverage to 60% of the population. The effective coverage. Health disparities are, in nicable disease prevention policies. private sector covers approximately six part, a consequence of the fragmentation Competing interests: None. million people, of which four million of resource pools and poor redistribution This article is part of a series commissioned by The come from obras sociales, contracting mechanisms. The national average BMJ based on an idea from the Harvard Global Health Institute. The BMJ retained full editorial control over private supplementary plans to about 200 infant death rate, for example, is 9.7 per editing and publication. 1 prepaid plans. To harmonise coverage 1000 of the population. But there is a Correspondence to: among funds, there is a mandatory health twofold difference between poor and rich [email protected] @RubinsteinOk the bmj | BMJ 2019;366:l5439 | doi: 10.1136/bmj.l5439 39 UNIVERSAL HEALTH COVERAGE

1 Cetrángolo O, Goldschmit A. Organización y 4 Instituto Nacional de Estadistica y Censos. Elaboration 7 Starfield B, Shi L, Macinko J. Contribution of financiamiento de la provisión de salud por based on EPH second quarter 2017-INDEC. https:// primary care to health systems and health. Milbank parte de la Seguridad Social en Argentina. www.indec.gob.ar/bases-de-datos.asp Q 2005;83:457-502. doi:10.1111/j.1468- Documento de Trabajo IIEP (UBA CONICET). 2018. 5 Rubinstein A, Zerbino MC, Cejas C, Lopez AS. 0009.2005.00409.x http://157.92.136.59/download/docin/docin_ Making universal healthcare effective in Argentina: a 8 Macinko J, Harris MJ. Brazil’s family health iiep_025.pdf blueprint for reform. Health Syst Reform 2018;4:203- strategy—delivering community-based primary 2 United Nations. Sustainable development goals. 13. doi:10.1080/23288604.2018.1477537 care in a universal health system. N Engl J https://www.un.org/sustainabledevelopment/ 6 Dirección de Estadísticas e Información en Salud. Med 2015;372:2177-81. doi:10.1056/ sustainable-development-goals/ Estadísticas vitales. Información básica Argentina NEJMp1501140 3 World Health Organization. Health systems financing - Año 2016. Buenos Aires: Ministerio de Salud de the path to universal coverage. World health report, la Nación; 2017. http://www.deis.msal.gov.ar/wp- Cite this as: BMJ 2019;366:l5439 2010. 2010. https://www.who.int/whr/2010/en/ content/uploads/2016/12/Serie5Numero59.pdf http://dx.doi.org/10.1136/bmj.l5439

40 doi: 10.1136/bmj.l5439 | BMJ 2019;366:l5439 | the bmj UNIVERSAL HEALTH COVERAGE

Investing in community health workers accelerates universal health coverage Wilhelmina Jallah and colleagues discuss Liberia’s successful community health assistant programme

llness is universal, access to health- ministry revised its National Community Initiative (IFI). This is part of the global care is not. The 71st World Health Health Services Policy.5 We decided that Integrating Community Health programme Assembly will be the first since community health workers should be led by USAID, Unicef, and the Bill and the World Bank and World Health paid and that they should be managed by Melinda Gates Foundation, and it monitors Organization released findings clinic based supervisors, such as nurses. the strengths and weaknesses of the Ishowing that over half of the world’s 7.3 We decided to increase the numbers of community health assistant programme.12 billion people, including a billion in remote community health workers in “last mile” The IFI focuses on key metrics such as rural communities, lack access to essential communities more than 5 km away from timeliness of community health assistant health services.1 These include prenatal a clinic and that they should be trained payments, restocking of supplies, quality care, vaccinations, and malaria treatment. more comprehensively in evidence based of clinical supervision, and community Compounding this crisis is a massive short- services. Finally, we strengthened the health assistant competency. Each age of healthcare workers, which is forecast systems supporting them, including quarter, partners convene to review these to grow to 18 million by 2030.2 the supply chain, logistics, and health data, translating new insights into quality Investing in community health workers, information. improvement activities. The IFI has inspired alongside more nurses and doctors, In July 2016, Liberia’s ministry of health initiatives to strengthen the supply chain can help close this gap. Expanding the led a coalition of non-governmental and use mobile technology to enhance coverage of evidence based healthcare organisations and donors to launch the training for community health assistants interventions through community health National Community Health Assistant and supervisors, for example. Ultimately, workers in the 73 countries with 97% of Programme. That coalition consisted of the IFI system improves accountability and the world’s preventable maternal and child organisations including the International adaptability, helping to maximise return on deaths could save at least 2.5 million lives Rescue Committee, Last Mile Health, investment in the programme. each year.3 But community health worker Partners in Health, Plan International, Paying and investing in community programmes vary widely in quality. Many the Clinton Health Access Initiative, and health workers isn’t just the right thing to don’t reach full potential because they lack funding partners including USAID, Unicef, do; it’s the smart thing to do to accelerate proper investment—especially payment. the Global Financing Facility, the Global universal health coverage. Liberia’s The majority of community health workers Fund for AIDS, Tuberculosis, and Malaria, community health worker policy reforms are either volunteers who receive no pay or and leading philanthropists. echo successful experiences in Bangladesh, semi-compensated cadres.4 We have seen remarkable progress. As Brazil, Ethiopia, and Rwanda, among We have seen this first hand. Approxi­ of March 2018, Liberia has hired, trained, others. More countries should adopt mately 60% of rural Liberians lack and equipped nearly 3000 community similar reforms. Together, we can realise access to healthcare or live more than 5 health assistants and over 300 supervisors the health-for-all vision articulated in the km (one hour’s walk) from the nearest (primarily nurses) across 13 of 15 Declaration of Alma Ata 40 years ago. No clinic.5 In response, Liberia’s ministry of counties—reaching 70% of the goal. These one has to be left behind, if we’re willing to health initially trained a national cadre of health workers have carried out over go as far as it takes. community health volunteers. But, because 340 000 home visits; treated over 61 000 they were unpaid and under supported, childhood cases of pneumonia, malaria, or We are grateful to Mallika Raghavan and Siobhan many dropped out or were ineffective in diarrhoea; screened nearly 75 000 children Kelley for their support and contributions to this practice. Data showed very low coverage for malnutrition; and supported 30 000 article. rates of essential health services.6 In pregnancy visits. They’ve also identified Wilhelmina Jallah is the minister of health of the one study, the proportion of children in over 1700 potential infectious disease Republic of Liberia. rural areas receiving treatment for acute trigger events.8 At our 2017 national health Francis Kateh is the chief medical officer/deputy respiratory infection remained at 6.6%.7 conference, multiple counties reported minister of health of the Republic of Liberia. This inequality places us all at risk. The lack increases in children receiving treatment Raj Panjabi is the chief executive of Last Mile Health of paid, well supported community health for malaria, pneumonia, or diarrhoea of and assistant professor at Harvard Medical School. workers helped fuel the spread of the 2014- over 50%, and facility based deliveries in Competing interests: None declared. 16 Ebola epidemic from rural communities one remote district increased from 55% to 9-11 This article is part of a series commissioned by The into cities, claiming thousands of lives. 84%. BMJ based on an idea from the Harvard Global Health In 2015, as the Ebola epidemic was We continue to optimise the programme Institute. The BMJ retained full editorial control over brought under control, we enacted bold by focusing on quality. We are developing editing and publication. reforms. In addition to keeping primary a national performance management Correspondence to: R Panjabi healthcare free of charge, Liberia’s health system called the Implementation Fidelity [email protected] @lastmilehealth the bmj | BMJ 2019;366:l5471 | doi: 10.1136/bmj.l5471 41 UNIVERSAL HEALTH COVERAGE

1 World Bank. Tracking universal health coverage: sites/default/files/documents/1864/USAID_FAH_ Health Conference. December 7, 2017. Ministry of 2017 global monitoring report (English). 2017. Report_digital_version_nov21-508.pdf. Health, Republic of Liberia. http://documents.worldbank.org/curated/ 5 Revised national community health services policy 10 White E, Downey J, Sathananthan V, et al. Effect of a en/640121513095868125/Tracking-universal- 2016-2021. Ministry of Health and Social Welfare, community health worker intervention on childhood health-coverage-2017-global-monitoring-report. Republic of Liberia. disease treatment coverage in rural Liberia: a 2 Limb M. World will lack 18 million health 6 Liberia Institute of Statistics and Geo-Information difference-in-differences analysis. Am J Public Health workers by 2030 without adequate investment, Services. Liberia demographic and health survey. [forthcoming]. warns UN. BMJ 2016;354:i5169. doi:10.1136/ 2013. https://dhsprogram.com/pubs/pdf/FR291/ 11 Luckow PW, Kenny A, White E, et al. Implementation bmj.i5169 FR291.pdf research on community health workers’ provision 3 Perry HB, Zulliger R, Rogers MM. Community 7 Kenny A, Basu G, Ballard M, et al. Remoteness of maternal and child health services in rural health workers in low-, middle-, and high-income and maternal and child health service utilization Liberia. Bull World Health Organ 2017;95:113-20. countries: an overview of their history, recent in rural Liberia: A population-based survey. J doi:10.2471/BLT.16.175513 evolution, and current effectiveness. Annu Rev Public Glob Health 2015;5:020401. doi:10.7189/ 12 Integrating Community Health Program Summary. Health 2014;35:399-421. doi:10.1146/annurev- jogh.05.020401 USAID/Unicef www.usaid.gov/what-we-do/global- publhealth-032013-182354 8 Community Based Information System. March 2018. health/cross-cutting-areas/integrating-community- 4 Gichaga A, Milestone D, Sharma P. Strengthening Ministry of Health and Social Welfare, Republic of health-program. primary healthcare through community health Liberia. workers: closing the $2 billion gap. Financing 9 Boima T. Overview of the National Community Health Cite this as: BMJ 2019;366:l5471 Alliance for Health, USAID. https://www.usaid.gov/ Assistant Program. Presentation at 2017 National http://dx.doi.org/10.1136/bmj.l5471

42 doi: 10.1136/bmj.l5471 | BMJ 2019;366:l5471 | the bmj UNIVERSAL HEALTH COVERAGE

Will artificial intelligence help universalise healthcare? Satchit Balsari explains how artificial intelligence has the potential to deliver healthcare to the billions of people who currently lack access

ealth and wellness centres practitioners are left prescribing placebo do harm. Low resource settings can are opening across India— combinations of anti-inflammatories, hardly afford to burden overstretched soon there will be 150 000 of analgesics, and antibiotics in lieu of providers with the responsibility of them.1 As part of the govern- meaningful care.5 6 feeding algorithms. The clinician as ment of India’s ambitious goal Virtual assistants, predictive algorithms, data entry operator ought to become Hto extend insurance coverage to 500 mil- customised care plans, and home based as anachronistic as the stenographer. lion citizens, thousands of rural and urban nudges (interventions that change Companies have begun to leverage clinics are being transformed. I visited one behaviour), all powered by artificial voice recognition technologies, natural such clinic in July, not far from Bengaluru, intelligence (AI), could offer a solution for language processing, and deep neural India’s IT capital. We were greeted by freshly delivering care to those with little to no networks to extract texts from physician- painted walls covered with symbols depict- access to healthcare.7 But for this vision to patient interactions to populate electronic ing various services offered at the clinic. be rolled out globally—to reach the billions health records.12 13 Making these advances The telemedicine screen in one such sym- of people who do not have the most basic accessible in local languages and dialects, bol hinted at the central role that digital ser- access to care—we need more than Silicon to liberate hundreds of thousands vices will play. The “mid-level healthcare Valley hubris and Wall Street speculation. healthcare workers from their keyboards, provider” staffing the clinic, was a member Deep neural networks—the complex, might be one of the most important of a new cadre of non-physician providers, multilayered, self teaching models that contributions that AI can make towards continuing India’s long tradition of suc- are expected to eventually aid and nudge expanding access. cessfully shifting the delivery of basic care physicians—will only be as good as the We need a thoughtful reconnaissance to community health workers.2 data that power them. The more structured in the global south, of what problems This combination of task shifting and and labelled the data, the easier it is need solving (first), what data are needed technology is a recognisable sign of primary for machines to ingest them. The most to solve them, and how best technology healthcare delivery anywhere in the global successful examples of AI being integrated can be leveraged to collect these data. south. Providers at such clinics across into healthcare are in areas where the data Clinicians should be at the forefront of Asia and Africa continue to see untenable are the cleanest, such as radiology images this rapidly changing landscape steering numbers of patients, with each interaction or pathology slides.8 Algorithms attempting developers and investors towards solving lasting no more than a few minutes.3 This to predict survival, re-admission to the most basic yet pressing challenges of fleeting visit, sometimes at the cost of hospital, or risks for infections are messier care delivery today. a day’s wage, is the patient’s only hope and highly dependent on the quality of data Satchit Balsari is assistant professor in emergency of being correctly screened, receiving a (and their accompanying biases) captured medicine at Beth Israel Deaconess Medical Center and 9 Harvard Medical School and is Burke Fellow at the diagnosis or treatment, or being referred. in electronic health records. Harvard Global Health Institute. He directs the India The success of universal health coverage Clinicians are familiar with the problems Digital Health Net, an interdisciplinary research and depends on the optimisation of care in of these context laden algorithms. Clinical policy collaborative at the Harvard Mittal Institute. these interactions. parameters considered “normal” in Competing interests: Received the Harvard Global Frontline providers are faced with European men should not have driven Health Institute Burke Fellowship. state mandated data entry requirements, clinical decision making around the world, This article is part of a series commissioned by The using only tablet computers and few as they did for as many decades.10 Similarly, BMJ based on an idea from the Harvard Global Health other resources at their disposal. Despite successful clinical trials often cannot be Institute. The BMJ retained full editorial control over editing and publication. the growth of point-of-care devices and replicated when validation is attempted Correspondence to: [email protected] teleconsulting services, few governments in entirely different populations. Services @HarvardGH have managed to test and successfully whose deep neural networks have been integrate portable technologies into powered by data rich populations might not 1 NITI Aayog. Strategy for New India @ 75. 2018. https://www.niti.gov.in/writereaddata/files/Strategy_ longstanding clinical pathways. Even a be applicable or useable in low and middle for_New_India.pdf 11 reliable list of current diagnoses, drugs, and income settings. Many applications might 2 Health and Wellness Centres under Ayushman laboratory results—information that most transcend population heterogeneity, but Bharat. https://pib.gov.in/newsite/PrintRelease. aspx?relid=188246. clinicians would consider very basic—is those that will be used to decide, deny, or 3 Irving G, Neves AL, Dambha-Miller H, et al. seldom available. That most patients and delay care must have contextual intelligence International variations in primary care physician physicians, in most countries, cannot to be relevant and fair. Where technology is consultation time: a systematic review of access basic clinical data collated from most needed, the data do not exist. 67 countries. BMJ Open 2017;7:e017902. 4 doi:10.1136/bmjopen-2017-017902 disparate sources is remarkable. With A drive to promote digitisation over 4 Balsari S, Fortenko A, Blaya JA, et al. Reimagining limited time, and even less information, other aspects of care will, however, health data exchange: an application programming the bmj | BMJ 2019;366:l5503 | doi: 10.1136/bmj.l5503 43 UNIVERSAL HEALTH COVERAGE

interface-enabled roadmap for India. J Med Internet Rev Cancer 2018;18:500-10. doi:10.1038/s41568- 12 Geraci J, Wilansky P, de Luca V, Roy A, Kennedy Res 2018;20:e10725. doi:10.2196/10725 018-0016-5 JL, Strauss J. Applying deep neural networks to 5 Hicks LA, Taylor THJr, Hunkler RJUS. US outpatient 9 Miller DD. The medical AI insurgency: what physicians unstructured text notes in electronic medical records antibiotic prescribing, 2010. N Engl J Med must know about data to practice with intelligent for phenotyping youth depression. Evid Based Ment 2013;368:1461-2. doi:10.1056/NEJMc1212055 machines. NPJ Digit Med 2019;2:62. doi:10.1038/ Health 2017;20:83-7. doi:10.1136/eb-2017- 6 Butler CC, Rollnick S, Pill R, Maggs-Rapport F, Stott N. s41746-019-0138-5 102688 Understanding the culture of prescribing: qualitative 10 Clinical trials have far too little racial and ethnic 13 Liao KP, Cai T, Savova GK, et al. Development of study of general practitioners’ and patients’ diversity. Scientific American. https://www. phenotype algorithms using electronic medical perceptions of antibiotics for sore throats. BMJ scientificamerican.com/article/clinical-trials-have-far- records and incorporating natural language 1998;317:637-42. doi:10.1136/bmj.317.7159.637 too-little-racial-and-ethnic-diversity/. processing. BMJ 2015;350:h1885. doi:10.1136/ 7 Topol E. Introduction to Deep Medicine. In: Deep 11 Sarah E, Loe JD, Robertson C. Who’s left out of big bmj.h1885 Medicine. Basic Books, 2019: 1-22. data. In: Cohen G, Lynch HF, Vayena E, Gasser U, 8 Hosny A, Parmar C, Quackenbush J, Schwartz LH, eds. Big Data, Health Law and Bioethics. Cambridge Cite this as: BMJ 2019;366:l5503 Aerts HJWL. Artificial intelligence in radiology. Nat University Press, 2018. http://dx.doi.org/10.1136/bmj.l5503

44 doi: 10.1136/bmj.l5503 | BMJ 2019;366:l5503 | the bmj