Article Text

Protocol
Protocol for a controlled before-after quasi-experimental study to evaluate the effectiveness of a multi-component intervention to reduce gaps in hypertension care and control in low-income communes of Medellin, Colombia
  1. Esteban Augusto Londoño Agudelo1,2,3,
  2. Tullia Battaglioli4,
  3. Alonso Soto5,6,
  4. José Vásquez Gómez7,
  5. Hernán Aguilar Ramírez7,
  6. Viviana Pérez Ospina8,
  7. Armando Rodríguez Salvá9,
  8. Patricia Ortiz Solórzano10,
  9. Dennis Pérez11,
  10. Rubén Gómez-Arias1,3,
  11. Patrick Van Der Stuyft2
  1. 1Facultad Nacional de Salud Pública, Grupo de Epidemiología, Universidad de Antioquia, Medellín, Colombia
  2. 2Faculty of Medicine and Health Sciences, Department of Public Health and Primary Care, Ghent University, Gent, Belgium
  3. 3Facultad de Medicina, Universidad CES, Medellin, Colombia
  4. 4Consultant, Basel, Switzerland
  5. 5Instituto de Investigación en Ciencias Biomédicas, Universidad Ricardo Palma, Lima, Perú
  6. 6Departamento de Medicina, Hospital Nacional Hipólito Unanue, Lima, Perú
  7. 7UPSS Santa Cruz, Metrosalud E.S.E, Medellin, Colombia
  8. 8Dirección Ejecutiva, PSICOL -Psicología Ocupacional S.A.S, Medellin, Colombia
  9. 9Centro de Epidemiología y Salud Ambiental, Instituto Nacional de Higiene, Epidemiología y Microbiología (INHEM), La Habana, Cuba
  10. 10Facultad de Medicina, Pontificia Universidad Católica del Ecuador, Quito, Ecuador
  11. 11Departamento de Vigilancia e Investigaciones Epidemiológicas, Instituto de Medicina Tropical Pedro Kourí, La Habana, Cuba
  1. Correspondence to Dr Esteban Augusto Londoño Agudelo; eaugusto.londono{at}udea.edu.co

Abstract

Introduction Research on public health interventions to improve hypertension care and control in low-income and middle-income countries remains scarce. This study aims to evaluate the effectiveness and assess the process and fidelity of implementation of a multi-component intervention to reduce the gaps in hypertension care and control at a population level in low-income communes of Medellin, Colombia.

Methods and analysis A multi-component intervention was designed based on international guidelines, cross-sectional population survey results and consultation with the community and institutional stakeholders. Three main intervention components integrate activities related to (1) health services redesign, (2) clinical staff training and (3) patient and community engagement. The effectiveness of the intervention will be evaluated in a controlled before-after quasi-experimental study, with two deprived communes of the city selected as intervention and control arms. We will conduct a baseline and an endline survey 2 years after the start of the intervention. The primary outcomes will be the gaps in hypertension diagnosis, treatment, follow-up and control. Effectiveness will be evaluated with the difference-in-difference measures. Generalised estimation equation models will be fitted considering the clustered nature of data and adjusting for potential confounding variables. The implementation process will be studied with mixed methods. Implementation fidelity will be documented to assess to which degree the intervention components were implemented as intended.

Ethics and dissemination The study protocol has been approved by the Ethics Research Committee of Metrosalud in Colombia (reference 1400/5.2), the Medical Ethics Committee of the Antwerp University Hospital (reference 18/40/424) and the Institutional Review Board of the Antwerp Institute of Tropical Medicine (reference 1294/19). We will share and discuss the study results with the community, institutional stakeholders and national health policymakers. We will publish them in national and international peer-reviewed scientific journals.

Trial registration number NCT05011838.

  • Hypertension
  • Organisation of health services
  • PUBLIC HEALTH
  • PREVENTIVE MEDICINE
  • PRIMARY CARE
http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • We respond to exhortations to rigorously evaluate multi-component interventions to improve hypertension care and control in low-income and middle-income countries.

  • The planned intervention is based on formative research and inspired by guidelines from international hypertension care initiatives and has been adapted to the local environment with the participation of all relevant stakeholders.

  • We employ a controlled before-after quasi-experimental study design.

  • We will rely on quantitative outcomes and qualitative process assessments, and combine effectiveness and implementation fidelity indicators to better inform subsequent adaptation and scaling up.

  • The intervention will be implemented in a low-income area of an industrialised city in a middle-income country, and results may not be generalisable to upper-class urban zones or underserved rural settings.

Introduction

Non-communicable diseases accounted for 73.5% of global deaths in 2017.1 Their incidence is disproportionately increasing in low/middle-income countries (LMICs),2 and they constitute the first cause of mortality and disability in the Americas region, generating at least 5.5 million deaths each year.3 Notwithstanding, most Latin American and Caribbean health systems provide inadequate chronic disease management.4 Globally, only about one in ten people with chronic conditions are successfully treated.5 Moreover, most out-of-pocket health payments and catastrophic expenditures are related to such conditions.6 7

Uncontrolled hypertension is the main modifiable risk factor for cardiovascular diseases (CVDs) and is associated with more than 10 million deaths worldwide each year.8 Almost 75% of people suffering from hypertension live in LMICs, where disease awareness is low, and healthcare access is limited.9 10 High disease prevalence and poor hypertension control are pivotal factors in the rising epidemic of CVDs in LMICs,11 where they mainly strike the working-age population and further hamper economic growth and social development.10 Among the 1.6 million people who yearly die from CVDs in Latin America and the Caribbean, half a million are below 70 years old.12 Nevertheless, limited research in this field has been performed in LMICs.13

A multi-country cross-sectional study conducted between 2003 and 2009, including 14 LMICs,14 found a prevalence of hypertension of 27.7% in adults, with only 46.5% and 32.5% of hypertension awareness and control, respectively, among the affected. A 2019 review focusing on LMICs15 pooled individual-level hypertension data on people aged 15 years and older from 44 countries and found a 17.5% hypertension prevalence. Of the hypertensive individuals, 39.2% were aware of their condition, 29.9% received pharmacological treatment and barely 10% attained controlled blood pressure (BP). These reports provide information on hypertension control internationally and on variations among populations. Most studies assessing patient and healthcare provider barriers to hypertension care have been conducted at the health facility level and in high-income countries.13 Research at the population level to document in-depth and to tackle the main gaps and barriers hampering equitable and effective hypertension care and control in LMICs is needed.

The Chronic Care Model16 and its expanded version5 aim to guide interventions for improving chronic disease management. It is the basis of the Standardised Hypertension Treatment and Prevention (SHTP) Project,17 developed by the Pan American Health Organization and the Centres for Disease Control and Prevention. The SHTP Project identified six key elements to strengthen health systems at the primary care level and to improve hypertension control, using evidence-based components: guideline-based standardised treatment protocols; effective drug procurement mechanisms using a core set of medications; registries for cohort monitoring and evaluation; patient empowerment; team-based care system and community engagement.17 This approach was reinforced by the WHO Global Hearts Initiative, primarily through the Technical Package for Cardiovascular Disease Management in Primary Health Care.18 Operational research to assess the impact of this and other similar strategies is urgently advocated.

The objective of the present study is to evaluate the effectiveness of a multi-component intervention to reduce the gaps in hypertension care and control in low-income communes of Medellin, Colombia. In addition, we aim to assess the process and fidelity of the intervention’s implementation.

Rationale

In 2016, we conducted a cross-sectional survey in a low-income commune of Medellin19 to estimate the prevalence of hypertension in the adult population and the magnitude and determinants of the main gaps in hypertension care and control. The definitions of those gaps are shown in table 1. In summary, we found a hypertension prevalence of 43.5%, and 28.2% aware and 15.3% unaware hypertensive individuals. Moreover, among the aware persons, 14.4% presented a follow-up gap, 93.4% were prescribed antihypertensive drugs, 38.9% were not compliant and 39.0% were uncontrolled. The survey results, the—hypothesised—possible determinants of the identified gaps (table 2), and potential solutions were discussed with communities, service providers and institutional stakeholders in a series of workshops. Based on the workshops’ conclusions, in consultation with the main stakeholders and inspired by the SHTP Project and the WHO Global Hearts Initiative,17 18 we designed a multi-component intervention to improve hypertension care and control in Medellin.

Table 1

Main gaps in hypertension care and control—definitions

Table 2

Potential determinants at the health provider, population and health system level of the main gaps in hypertension care and control

Methods and analysis

Overview of study design

The effectiveness of a multi-component intervention in reducing gaps in hypertension care and control will be evaluated in a controlled before-after quasi-experimental study. A graphical summary of the underlying theory of action is depicted in figure 1. Besides outcomes, the process and implementation fidelity will be studied to assess to which degree the intervention components were implemented as intended. The study will use mixed methods, comprising a baseline and an endline population survey, medical records and healthcare registers analysis, self-registration forms of intervention activities, non-participant observations and interviews with key informants such as healthcare staff, managers and patients and their caregivers. The intervention will start in the first quarter of 2023 and be evaluated after 24 months.

Figure 1

Model conceptualising the theory of action of a multi-component intervention to improve hypertension care and control in low-income Medellin, Colombia.

We checked the completeness and accuracy of our research protocol using the Standards for Quality Improvement Reporting Excellence reporting guidelines (online supplemental file 1).

Patient and public involvement

From the onset of the planning phase of the intervention onwards, the main community and institutional stakeholders—the local patient’s association, the community representatives to the city council, managers and clinical staff of the city’s public health provider (Metrosalud) and public health practitioners from the University of Antioquia—were invited for and participated in a series of workshops to discuss the results of the previous formative research. Hence, patients, community and institutional stakeholders actively proposed pathways to improve hypertension care and control within the community and at the health facility level. We will continuously further involve all local stakeholders in monitoring, analysing and learning from the forthcoming intervention.

Study setting and population

Colombia is a middle-income country with a 2019 per-capita gross domestic product of US$6429 (current US$).20 The mortality from CVDs in 2017 was 150.3 per 100.000 population, and 30.5% of all deaths and 16.7% of years of life lost can be attributed to CVDs.21 The Colombian health system has two different health insurance schemes run by health insurance companies. People able to contribute and their beneficiaries are compulsorily affiliated to the contributory scheme, which covers the formally employed workers, pensioners and part of the self-employed. The state finances the subsidised scheme, which protects people who cannot afford contributions.

Medellin is the second-largest industrialised city in Colombia. It has 16 communes, and its total 2016 population was around 2.5 million.22 The intervention will be implemented in Commune 2, northeast of Medellin. Commune 6, located northwest of Medellin, was selected as the control area to deliver routine care. Both communes are among the most deprived areas of the city23 and share similar characteristics: urban setting, low-income population, typical functioning of the national health system and commitment of Metrosalud—the city’s public healthcare provider—to develop improvement programmes based on research results.

Intervention

A multi-component intervention to improve hypertension care and control will integrate activities related to (1) health services redesign, (2) clinical staff training and (3) patient and community engagement. Each intervention component will be standardised, and its related activities described in detail in a field manual. The intervention will be implemented by health services staff with technical assistance from the investigators. The three components of the intervention are detailed in the following sections.

Health services redesign

‘Healthy Hearts’ service

A nursing station will be established to provide BP measurement, cardiovascular risk assessment and preventive counselling during extended opening hours. It will secure hypertensive patients’ effective follow-up, especially for presumptive hypertensive individuals and those newly diagnosed. The service will facilitate serial, automated, self-measured BP testing to confirm diagnosis within the health facility premises in order to exclude white coat hypertension. It also will administer periodical anonymous exit questionnaires to patients that attended hypertension-related consultations to assess the quality of care. Results will be communicated to the cardiovascular risk team for analysis.

Hypertension screening

The healthcare flow for patients with a high BP measurement will be standardised using a written procedure. All adults attending health services who did not have their BP measured by a healthcare provider in the previous year will be referred to the Healthy Hearts service for screening. Furthermore, doctors and nurses fortuitously encountering presumptive hypertensive individuals will refer them to the Healthy Hearts service to confirm the diagnosis.

Clinical management
Creation of the cardiovascular risk team

A group of doctors supervising clinical hypertension management and coordinating action plans for improvement. It will conduct audits, periodically selecting a representative sample of clinical records of hypertensive patients and use a structured questionnaire to evaluate the fulfilment of the requirements of a complete cardiovascular risk assessment interview and physical examination, doctors’ adherence to guidelines and the quality of the provided care. It also will supervise the Healthy Hearts service. This operational body will hold weekly meetings and provide regular feedback and support to health professionals involved in hypertension care.

Guideline-based standardised diagnostic and treatment protocols

A simplified diagnostic and treatment algorithm based on the national clinical guidelines and in line with the Global Hearts Initiative will be agreed on with the general direction of Metrosalud, the direction of the health area and the clinical staff. It will identify a core set of primary and secondary options within each of the major classes of antihypertensive medication.

Availability of antihypertensive medications

It will be assured through procurement mechanisms agreed with the central drug store of Metrosalud. To timely inform on the possible need for alternative prescriptions, the availability of antihypertensive medications will be communicated to clinicians at the beginning of each week and ad hoc in case of stock-out.

Clinical staff training

Training on good clinical management of hypertension

It will be focused on correct BP measurement, use of evidence-based guidelines, cardiovascular risk assessment, a standardised diagnostic and treatment algorithm, correct prescription of pharmacological and non-pharmacological treatment, patient counselling and prevention of clinical inertia.

Training on communication skills and patients’ needs assessment

This training will be designed under the ‘patient-centred medicine’ framework,24 aiming to equip all health workers involved in hypertension care with tools for understanding patients’ feelings and experiences of illness and improving their capacity to address social, psychological and behavioural dimensions of hypertension care.

Patient and community engagement

Patients’ empowerment

‘Expert hypertensive patients’, under the supervision of a social worker, will provide support and transmit their know-how to other patients in need, particularly those newly diagnosed or non-adherent to treatment or presenting uncontrolled hypertension.

Community engagement

A Community Hypertension Outreach Group will be set up, composed of existing voluntary community health workers, who will be trained and certified. This group will conduct BP measurements in selected public areas of the commune, referring those with positive screening to the nearest health facility for diagnosis confirmation. It will also provide health information with an emphasis on healthy lifestyles. Additionally, local communication channels such as the community radio and the local newspaper will be engaged.

Effectiveness assessment

The primary outcome will be the control gap in the total hypertensive population (aware and unaware hypertensive individuals). The secondary outcomes will be the diagnostic, treatment, follow-up and control gaps defined in table 1.

Inclusion and exclusion criteria for the baseline and endline surveys

Eligible persons will be all 35 years or older permanent inhabitants of the selected communes, willing and able to provide written informed consent. Potential participants with a mental disability or unable to answer the questionnaire will be excluded.

Sample size and sampling procedures for the surveys

Expecting at baseline 60% non-control in the 35 years or older hypertensive population and at endline 10% difference in difference (baseline minus endline) between the intervention and control arms, a sample of 385 hypertensive individuals per arm is required for 80% power and a 95% confidence level. Given a 45% prevalence of hypertension in individuals aged 35 years or older, screening 856 individuals will allow finding the needed number of hypertensive individuals. We will increase the number screened to 1190 individuals, assuming 10% non-response and, based on previous observations, a design effect of 1.25 for our sampling scheme.

A cluster sample will be drawn from a sampling frame consisting of the addresses of all premises provided by the municipality and using maps of Medellin detailing house blocks. Given an average of 1.5 individuals ≥35 years old per household, 795 households will be sampled per study arm by selecting 53 clusters of 15 households from the different neighbourhoods of the communes, with an allocation of the number of clusters proportional to neighbourhood size. A group of trained interviewers will make up to two repeat door-to-door visits to every selected household to include all identified eligible household members aged 35 years or older in the surveys.

Data collection

Using a structured questionnaire, participants will be interviewed on sociodemographic characteristics, health-seeking behaviour, cardiovascular risk factors and previous hypertension diagnosis. In addition, participants aware of being hypertensive will be questioned on their follow-up, antihypertensive pharmacological and non-pharmacological treatment, and treatment compliance. BP will be measured using a digital manometer, following international recommendations for standardised BP measurement in population surveys.25 26

Operational definitions for the surveys

Hypertensive individual

Self-report of previously diagnosed hypertension or without a previous diagnosis but presenting an average repeat BP measurement equal to or higher than 140/90 mm Hg.14 27

Controlled hypertensive individual

Self-report of previously diagnosed hypertension and an average repeat BP measurement of less than 140/90 mm Hg for patients between 35 and 59 years old or diabetic patients and below 150/90 mm Hg for patients aged 60 years or older.28

Data analysis

The difference-in-difference measures29 30 will be calculated (endline-baseline changes in the intervention minus endline-baseline changes in the control commune) for the primary and secondary outcomes. Where relevant, we will stratify on affiliation to subsidised or contributory insurance schemes. Generalised estimation equation models will be fitted, considering the clustered nature of the data, and adjusting for potential confounding variables. Unadjusted and adjusted ORs and their 95% CI will be calculated. The Statistical Package for Social Sciences V.24 (SPSS Inc) will be used for data analysis.

Assessment of the implementation process and fidelity

The indicators to measure results at the health service level are shown in table 3. They were elaborated considering the performance indicators proposed by the SHTP Project, the Global Hearts Initiative and other propositions to improve the global prevention and control of CVDs.17 18 31 32 The indicators we constructed for monitoring implementation are listed in table 4. The needed data will be collected in the intervention and control areas, at baseline, during implementation and at the endline. Electronic clinical and billing records of Metrosalud, registers of the ‘Healthy Hearts’ service, exit interviews and clinical audits will be the data sources.

Table 3

Performance indicators at the health facility level

Table 4

Quantitative indicators for monitoring implementation

For assessing implementation fidelity,33 34 we will use self-registration forms for specific intervention activities, alongside non-participant observations and semi-structured key informant interviews. Self-registration forms will be collected monthly to determine whether actual implementation is done as intended. We will also select a heterogeneous sample of activities to be observed purposively and actors to be interviewed. Observations will provide real-time information on implementation, while interviews will throw light on observed practices and reasons for non-adherence. Additionally, overall experience with the intervention will be explored through in-depth interviews with stakeholders and patients. Data saturation will be considered to determine the frequency of observations and the number of interviews. A local social scientist will analyse data with NVivo V.10 (QSR International Pty LTD, Melbourne, Australia, 2010). A senior sociologist not involved in data collection will review the coding consistency.

We will determine whether intervention components were implemented as planned or were modified, deleted or added35 and assess adherence to the content and dose (ie, frequency, duration and coverage) of components and the influence of moderating factors such as comprehensiveness of intervention description, facilitation strategies, quality of delivery, recruitment, participant responsiveness and context. Factors influencing implementation will be classified using these pre-established categories, and aspects not matching this coding will be categorised as intervention-specific moderating factors.36 Finally, reasons for adapting intervention components will be identified through an inductive thematic analysis of key informants’ interview responses.

Finally, we will contrast outcome data with implementation data and triangulate quantitative and qualitative results obtained from different information sources to validate and complement our findings.

Data management

Databases will be electronically encrypted, and, when relating to individuals, a single study code will be assigned to each participant. The log of participant names and their assigned codes will be stored in a separate database with restricted access. For the surveys, questionnaires will be filled in employing tablets, and the information will be consolidated using an electronic platform. We will develop checks for data entry with built-in filters and logical constraints to assess the completeness and accuracy of our data. Recorded qualitative information will be transcribed verbatim by a local social scientist. Transcripts will be anonymised. Study-related material will be securely stored. Paper documents and tapes will be locked, and electronic databases and backup copies will be password protected. Data will be retained following local legislation.

Ethics and dissemination

The study will be conducted in accordance with the Declaration of Helsinki of research ethics. All field staff will be trained in responsible conduct of research. Informed consent will be obtained from survey participants and partakers in the qualitative research components (online supplemental file 2). A convenient place and time for interviews will be assured. Confidentiality will be guaranteed throughout. In particular, exit interviews will be anonymous and will not inquire into the identity of the health personnel the interviewee consulted with. Patients who are hypertensive in the survey and without follow-up or treatment will be referred for care to a provider of their health insurance scheme.

The planned intervention is based on formative research and inspired by guidelines from international hypertension care initiatives and has been adapted to the local environment with the participation of all relevant stakeholders. It does not pose health risks to beneficiaries as it implies implementing evidence-based best practices in the standard health setting. The local patients’ association, study site health staff and Metrosalud central authorities have agreed on any new patient care procedure. All activities are implemented conditional on adjustment when needed and planned to be continued beyond the study period.

The study protocol (online supplemental file 3) has been reviewed and approved by the Ethics Research Committee of Metrosalud in Colombia (reference 1400/5.2), the Medical Ethics Committee of the Antwerp University Hospital (reference 18/40/424) and the Institutional Review Board of the Antwerp Institute of Tropical Medicine (reference 1294/19).

The study findings will be shared and discussed in workshops with the local health authorities, health staff, representatives of Metrosalud and patient organisations in view of adjusting activities where needed, strengthening sustainability and preparing to scale up the intervention. Finally, results will be further disseminated through the media, presented at national and international conferences and published in peer-reviewed scientific journals.

Implications of this research

There is an urgent need, particularly in LMICs, to develop and test public health approaches that tackle the rising burden of non-communicable diseases.37 Doing so within an integrated framework of care will benefit all patients.5 When developing our proposal, we started with the international hypertension initiatives’ guidelines17 18 31 32 and stepped them up considering the local context. We will thus implement and evaluate a comprehensive multi-component strategy involving stakeholders from community members to health programme administrators, who all participated in the protocol’s design. We believe this ensures social relevance and will increase intervention uptake at the health service and community levels.

The quasi-experimental nature of our design will provide solid evidence regarding the outcome of the intervention, which can form the basis for further economic evaluation. Relevant lessons will also be learnt on what functioned, for whom and why within a specific national health system’s complex dynamics. While findings should not be extrapolated to upper-class urban zones or underserved rural areas, over half of the Colombian population lives in low-income urban environments, to which the results may be readily generalisable.

Our research closely involves Metrosalud, the most significant public healthcare provider of Medellin. This major stakeholder in the Colombian health system intends to further explore the effectiveness of the multi-component hypertension care and control intervention in other health areas of its network. Based on the results of the present trial, replication studies will be set up to evaluate and contrast, using routine data, the outcomes of the strategy in rural and urban settings and health areas with different socioeconomic levels. Study findings can also inform and inspire other initiatives for better chronic care in LMICs, particularly in Latin America. Finally, assessing the implementation fidelity of the intervention and its determinants may provide valuable insights into the scalability of the WHO Global Hearts Initiative.

Ethics statements

Patient consent for publication

Acknowledgments

The authors gratefully acknowledge the inputs towards designing the intervention of the different Colombian health sector stakeholders mentioned in the text. This study is part of a concerted research effort within the Latin American Network for Multidisciplinary Research on Chronic Diseases, a regional organisation of partner academic institutions supported by the Belgian Directorate-General for Development Cooperation.

References

Supplementary materials

Footnotes

  • EALA and TB contributed equally.

  • Contributors EALA and TB contributed equally to this paper. PVDS is guarantor for the overall content. PVDS, EALA, and TB conceived the study and took the lead in conceptualising and designing the intervention, with contributions from ARS, AS, RG-A and POS. JVG, HAR, VPO, RG-A and EALA coordinated different facets of the stakeholder involvement and engaged in the workshops in which the content of the intervention activities and operational procedures were defined. DP developed qualitative aspects of the process assessment and the evaluation of the implementation fidelity, with support of PVDS. All authors significantly contributed to refining the research protocol and the data collection tools. AS, EALA, PVDS and TB drafted the manuscript. All authors commented on subsequent versions of the paper and read and approved the final manuscript.

  • Funding The work received partial funding from the National School of Public Health of the University of Antioquia (Colombia). The funding organisation did not intervene in the study’s design, the decision to publish this protocol and the writing up of the manuscript.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.