Patient-reported outcomes are based on patient (or caregiver) descriptions without direct measurement by a health care provider. To capture patient-reported outcomes, various patient-reported outcome measures (PROMs) have been created. Using PROMs has been linked to improved patient satisfaction, patient-provider communication, and clinical outcomes in many pediatric fields. Despite a long-standing history of utilizing PROMs for the evaluation and management of childhood asthma, pediatric pulmonologists lag behind other pediatric subspecialists in the use of PROMs.
During the National Heart, Lung, and Blood Institute’s “Defining and Promoting Pediatric Pulmonary Health” workshop, critical knowledge gaps and research opportunities in the use of PROMs for childhood respiratory health were reviewed. In particular, PROMs can be employed as screening tools in the general population for the primary or secondary prevention of pediatric lung diseases. Incorporating these PROMs into the pediatric primary care setting would be especially impactful. In addition, the use of PROMs for the evaluation and management of asthma suggests that they can be applied to other childhood respiratory diseases. Ongoing multicenter studies or national consortia that study pediatric lung diseases could be leveraged to conduct research designed to develop, validate, and assess the utility of PROMs to assess childhood respiratory health. Harnessing the electronic health record will be critical for the successful adoption of PROMs in children with lung diseases. Ultimately, an integrative approach to systematically address numerous barriers at the level of the provider, patient, and health care system will be needed to attain this goal and achieve sustainability.
Patient-reported outcomes (PROs) are based on patient (or caregiver) descriptions without direct measurement by a health care provider. PROs are often broad, rarely observable, and may provide insights into the impact of a condition, treatment, or care on a patient’s daily symptoms, health-related quality of life (HRQOL), functional status, or overall life satisfaction.1,2 PROs parallel the shift in clinical medicine toward patient-centered medical care and allow for the assessment of factors beyond standard physiologic measurements, including those related to health beliefs and cultural realities. In pediatrics, the inclusion of a child’s perspective is essential and often provides a different window into the disease experience.
To capture PROs, various identifiable, valid, and reliable patient-reported outcome measures (PROMs) have been created. Using PROMs is a recommended part of routine care for many pediatric fields and clinical settings because this has been linked to improved patient satisfaction, patient-provider communication, and clinical outcomes.3–5 However, despite these known benefits, the uptake of PROMs in pediatrics is limited by multiple real-world barriers to their utilization in clinical practice.2
During the National Heart, Lung, and Blood Institute’s “Defining and Promoting Pediatric Pulmonary Health” workshop, critical knowledge gaps and research opportunities in the use of PROMs for childhood respiratory health were reviewed (Table 1). In particular, PROMs can be employed as screening tools in the general population for the primary or secondary prevention of pediatric lung diseases. PROMs can also be used for the evaluation and management of childhood respiratory diseases. To this end, we first examined how PROMs are being applied in childhood asthma, the most common chronic lung disease in the pediatric population. Second, we reviewed how PROMs are being developed in another pediatric subspecialty that also takes care of children with chronic diseases. Third, we considered how using a PROM that captures pediatric respiratory health could be adopted into the primary care setting. Last, we evaluated how technology, and specifically the electronic health record (her), can help to make PROMs more widely available and facilitate their use in the pediatric population.
Critical Knowledge Gaps and Research Opportunities for the Use of PROMs for Childhood Respiratory Health . |
---|
Determine the feasibility of designing and implementing PROMs that could be employed as a screening tool in the general population for the primary or secondary prevention of pediatric lung diseases. |
Select the optimal domains of childhood respiratory health that need to be captured with PROMs. |
Understand the numerous obstacles at different levels that can hinder the utilization of these PROMs in the primary care setting. |
Learn strategies for success from the use and research of PROMs in childhood asthma and in other pediatric subspecialties. |
Identify how the electronic health record and other technologies can help to make PROMs more widely available and facilitate their adoption in the pediatric population. |
Critical Knowledge Gaps and Research Opportunities for the Use of PROMs for Childhood Respiratory Health . |
---|
Determine the feasibility of designing and implementing PROMs that could be employed as a screening tool in the general population for the primary or secondary prevention of pediatric lung diseases. |
Select the optimal domains of childhood respiratory health that need to be captured with PROMs. |
Understand the numerous obstacles at different levels that can hinder the utilization of these PROMs in the primary care setting. |
Learn strategies for success from the use and research of PROMs in childhood asthma and in other pediatric subspecialties. |
Identify how the electronic health record and other technologies can help to make PROMs more widely available and facilitate their adoption in the pediatric population. |
Summary of Meeting
PROs in Childhood Asthma
Childhood asthma, a chronic lung condition interspersed with acute exacerbations, demonstrates the importance and value of PROMs in understanding disease burden, the progression of symptoms, and treatment responses. Childhood asthma is often associated with preserved or normal physiologic measures at baseline (such as lung function), making it difficult to demonstrate improvements in these measures with interventions, and thus creating an opportunity for the use of PROMs.
Several asthma-specific PROMs that can be used in children (or their caregivers) are available.1 These include the Asthma Control Questionnaire (ACQ), the Pediatric Asthma Quality of Life Questionnaire, the Pediatric Asthma Caregiver Quality of Life Questionnaire, the Childhood Asthma Control Test, and certain item banks from the Patient-Reported Outcomes Measurement Information System (PROMIS).6 These validated measures can be used in both the clinical and research settings to monitor disease activity and assess the efficacy of interventions. Likewise, many of these are easy to use, publicly available, can be employed longitudinally, can be administered through different methods (eg, in person, by phone, or by sending the PROMs through regular mail, e-mail, short text message, or a web platform), and have established minimal clinically important differences (Table 2).
Instrument . | Minimally Important Difference . | Score Range . |
---|---|---|
Childhood Asthma Control Test | 2 points | 0–27 |
Pediatric Asthma Quality of Life Questionnaire | 0.5 points | 1–7 |
Pediatric Asthma Caregiver’s Quality of Life Questionnaire | 0.5 points in a domain or overall | 1–7 |
Asthma Control Questionnaire | 0.4 (children), 0.5 (adults) | 0–6 |
Instrument . | Minimally Important Difference . | Score Range . |
---|---|---|
Childhood Asthma Control Test | 2 points | 0–27 |
Pediatric Asthma Quality of Life Questionnaire | 0.5 points | 1–7 |
Pediatric Asthma Caregiver’s Quality of Life Questionnaire | 0.5 points in a domain or overall | 1–7 |
Asthma Control Questionnaire | 0.4 (children), 0.5 (adults) | 0–6 |
Each asthma-specific PROM generally provides a unique measure of disease activity and/or its impact. In addition, some of these also capture relevant disease-specific events, such as episodes of poor asthma control or exacerbations. Nguyen and colleagues examined the interrelationships of the ACQ with other self-completed PROMs (including the Childhood Asthma Control Test, Pediatric Asthma Quality of Life Questionnaire, and the Asthma Symptom Utility Index) among 305 children aged 6 to 17 years who were enrolled in a trial of add-on proton pump inhibitors in uncontrolled asthma.7 The authors reported good, but not perfect, correlations between the ACQ and the other PROMs (with correlation coefficients ranging from −0.64 to −0.73), along with variations in correlation based on the age of the child. In addition, the authors demonstrated greater ACQ scores (indicating worse asthma control) in children who experienced episodes of poor asthma control (as defined by a decrease in peak expiratory flow rate, urgent care visits, use of oral corticosteroids, or increased use of rescue medication) than in those who did not experience these episodes.
The use of PROMs can also effectively supplement physiologic measures of asthma control, such as the forced expiratory volume in 1 second, particularly because these are generally in the normal range for most children with mild-to-moderate asthma. For example, among 1746 respondents aged 12 years and older in the Observational Study of Asthma Control and Outcomes study, there was a weak negative and nonsignificant association (correlation coefficient = −0.17) between ACQ-5/ACQ-6 scores and forced expiratory volume in 1 second percent predicted. The authors attributed this to a high degree of unexplained variation between these measures, emphasizing the importance of measuring both symptom-related PROs and lung function to provide the broadest assessment of asthma activity, severity, and impact.8 The need to concomitantly measure PROs and lung function is highlighted in current asthma guidelines that recommend employing both for the evaluation and management of asthma in children.9
Use and Ongoing Research of PROMs in Other Pediatric Subspecialties
The extensive use of PROMs for the evaluation and management of asthma suggests that PROMs can be applied to other childhood respiratory diseases. Pediatric nephrology is a subspecialty that has incorporated the use of PROMs in the evaluation of chronic diseases that often interfere with multiple aspects of a child’s daily life. Thus, pediatric nephrologists are more likely to consider PROs as an important aspect of a child’s patient-centered medical care.2 In addition, the development, validation, and assessment of the utility of subspecialty- and condition-specific PROMs is an active area of research in pediatric nephrology.10
One example of this is the MyKidneyHealth study, which capitalizes on the ongoing Chronic Kidney Disease in Children (CKiD) study, a large, ongoing, multicenter, observational, prospective cohort of children, adolescents, and young adults with a history of chronic kidney disease (CKD). Established in 2003 by the US National Institute of Diabetes and Digestive and Kidney Diseases, the authors of the CKiD study aim to identify risk factors for the progression of CKD.11 The CKiD study has enrolled ∼900 children from >50 pediatric nephrology centers in the United States and Canada. In 2015, investigators obtained additional funding through the US National Institute of Arthritis and Musculoskeletal and Skin Diseases to recruit ∼200 children aged 8 to 17 years enrolled in the CKiD study and their caregivers with the goal of examining whether fatigue- and sleep-related items from PROMIS6 are clinically relevant and valid in children with CKD (NIAMS U19-AR069525) because fatigue and sleep issues are common in children with CKD and have been shown to be inversely related to HRQOL.12
For the MyKidneyHealth study, the process started with a rigorous validation phase that was designed to identify which PROMIS fatigue- and sleep-related items were salient to children with CKD, their caregivers, and their health care providers. First, 20 expert clinicians were presented with the PROMIS Pediatric Fatigue, Sleep Disturbance, and Sleep-Related Impairment item banks. Through a card sort activity, they were then asked to rank the top 4 items that they would ask a child with CKD to better understand the child’s level of fatigue and sleep health. Next, a group of children with CKD and their caregivers participated in similar card sort activities, rating items independently of each other. In addition, a subgroup of these children with CKD and caregivers participated in one-on-one cognitive interviews to assess whether they understood the items being ranked, with most of them demonstrating a full understanding of all measures. Trained research staff then conducted semi-structured one-on-one interviews with another subgroup of children with CKD and their caregivers to probe them about the impact of fatigue and sleep-related issues on their past and present daily life.13 These interviews were audio-recorded, transcribed, and coded for meaning units, which were defined as conceptually distinct statements in the words of the participant regarding their experiences of fatigue and sleep-related issues and the impact of these on their daily life.
Because the PROMIS Pediatric Fatigue, Sleep Disturbance, and Sleep-Related Impairment questions were developed with input from children from the general population, deductive content analysis was used to contrast the reports of symptoms from children with CKD and the impact on their daily life. In general, children with CKD reported similar breadth and type of lived experiences as children drawn from the general population. The final items applied to the whole MyKidneyHealth study population were selected according to the results of the aforementioned steps by considering (1) the proportion of children with CKD, their caregivers, and clinicians who identified the concept as important or relevant, (2) psychometric properties (statistics that indicated high precision and coverage at the high end of fatigue and sleep), (3) the results of cognitive interviews (if questions were understandable), and (4) balance across subdomains (eg, balance across impact on physical, psychological, and social functioning). Ultimately, 6 child-reported outcome measures (fatigue, sleep disturbance, sleep-related impairment, life satisfaction, meaning and purpose, and psychological experiences of stress) and 5 caregiver-reported outcome measures (pediatric global health, family relationships, positive affect, depressive symptoms, and anxiety) were selected. Each set of measures included 4 to 8 questions that included Likert-type responses and used a recall period of 7 days, with a total of 36 questions for the child and 23 for the caregiver. The results from the clinical validation phase supported that PROMIS items that focus on universally experienced fatigue- and sleep-related domains were relevant for children with CKD.
After the validation phase, all children with CKD and their caregivers were followed over 2 years and for a total of 7 time points to identify whether the fatigue- and sleep-related PROs being captured with the selected PROMIS items were stable over time and how these were impacted by sociodemographic and clinical characteristics. In total, the selected PROMIS items were completed at all 7 time points by ∼77% of children and ∼84% of caregivers enrolled. The findings from the MyKidneyHealth study will help to better understand how fatigue- and sleep-related PROs change over time in children with CKD and whether changes in these PROs correlate with changes in kidney function or other clinical parameters. In addition, although sleep-related issues are frequently cited in the literature as commonly experienced symptoms among children with CKD, pediatric nephrologists do not routinely assess sleep health during clinic visits.
If the MyKidneyHealth study demonstrates that sleep-related issues are indeed relevant for children with CKD, the selected PROMIS items present an opportunity for a quick and reliable assessment of this domain that can be integrated into subspecialty care and open patient-health care provider conversations about sleep health. Whereas the PROMs used in childhood asthma have been extensively validated, this vigorous validation process will be essential for the application of PROMs to other pediatric respiratory diseases.
PROs for Pediatric Respiratory Health: Translation to Primary Care
PROMs are not routinely used in the pediatric primary care setting; however, there is experience among community-based practices using comprehensive, age-specific, validated screening tools in the pediatric population. Specifically, the Bright Futures initiative from the American Academy of Pediatrics includes 6 areas as screening tools in all children from birth to age 21 years.14 Unfortunately, evidence reveals that recommendations for questionnaire-based screenings are not universally applied in primary care. The reasons for this are multifold but are mainly related to barriers at the level of the providers, patients, and health care system. To date, no subspecialty-specific PROMs have been employed as screening tools in primary care for the primary or secondary prevention of pediatric lung diseases.
In contrast, the PROMs used to assess pediatric asthma severity are perhaps the only subspecialty-specific PROMs routinely used in primary care. A deeper look at provider-level barriers reveals multiple limitations on the adaptation of other PROMs into primary care. These barriers include the provider’s limited time and resources (staff, space, and technology), fear of intrusion on practice, and skepticism over the clinical need of PROMs.2,15,16 Other considerations include the role of the provider’s adherence to guidelines, clinical knowledge, and language barriers.
Particular attention should be given to time as a potential barrier at the provider level because it is one of the most commonly cited obstacles to making a change in practice. Providers who see children in primary care are limited in the time they have to spend during a well-child visit and have a lot to accomplish in that short time. With such limited time, we must consider whether subspecialty- or condition-specific PROMs in the pediatric primary care setting make sense. In their work, Schonwald and colleagues make the important point that the perceived impediment of not having enough time to use screening tools should not prevent the adoption of a screening tool.17 It has also been noted that using a screening tool allows both the patient and the provider to focus the agenda for a visit. Thus, arguably, time could actually be saved if a visit starts with a better picture of the patient’s overall condition. In addition, although a little more than one-quarter of pediatric patients have a chronic condition, their care can be substantially fragmented. Gaining knowledge about a child or the family’s HRQOL through PROMs supports a comprehensive view that includes physical, emotional, social, and functional understanding for the provider.2 It is with this comprehensive understanding that providers are likely to gain insight into healthcare utilization data as well as positively impact the quality and value of their care.
The provider’s adherence to using recommended screening tools as part of guidelines is another crucial barrier to evaluating more closely. Screening tools are only as good as their adoption, and adoption can take considerable time. This is exemplified by the recommendations for the use of screening tools for developmental surveillance in the pediatric primary care setting, which were first made by the American Academy of Pediatrics in 2001. Yet, nearly a decade after the initial recommendations, on average only ∼50% of general pediatricians were adherent to these guidelines.18 Lipkin and colleagues have shown that the adoption of screening tools may be further impacted by implicit (unconscious) provider biases, which could lead to important health inequities in the care of children of certain backgrounds.18
In addition to provider-level barriers, patients aren’t routinely compliant with completing questionnaire-based screenings. Time, confidentiality, transportation, technology knowledge and access, accessible resources, and the provision of screening tools in a patient’s preferred language and at an appropriate health literacy level can be significant barriers.
Insurance companies can also create numerous obstacles to the use of PROMs at the community level. These include inconsistent or limited payments or shifting of costs to patients and their families who then refuse future screening. By providing adequate insurance payments, providers can fund staff and infrastructure dedicated to the adoption of screening tools.
Finally, understanding the practical process of how PROMs can be integrated into an existing clinic workflow must be considered. Providers can make major changes in response to a value proposition for patients. The acceptance of a screening tool is additionally promoted when it is easy to use, involves limited time to complete for the patient, and can be efficiently reviewed and scored by the provider.2 The Bright Futures periodicity schedule, with its broad, efficient, and simple recommendations for screening during routine preventive care, must be kept in mind when designing and determining the feasibility of translating a PROM to assess pediatric respiratory health in the primary care setting.
The Application of PROs in the EHR
Technology is key to promoting the efficiency of and enhancing the adoption of PROMs in all pediatric fields and clinical settings. In particular, the EHR is a critical component for the implementation of PROMs because it provides information technology infrastructure, maintains efficiency, and minimizes disruptions to workflow. EHRs must address these barriers while allowing the customization and flexibility needed to address the varying needs of practices, providers, and patients.
By offering multiple methods to distribute PROMs, EHRs can accommodate varied clinical practice needs and systems. PROM distribution in an EHR can be automated, which promotes efficiency. PROMs that are specific to a health care visit can be assigned in an automated fashion to patients based on the type of provider, appointment, subspecialty, division, or department. In addition, rules can identify a subset of patients with specific characteristics (eg, a child with a diagnosis of asthma) to facilitate a more targeted administration of PROMs. PROMs can be distributed manually, either to an individual patient or in bulk, through an EHR patient portal. Patient-initiated PROMs are those made available in an EHR patient portal that a patient can elect to complete at any time. Likewise, automatic rules can be created to send scheduled PROMs to a patient through an EHR patient portal, which is ideal for monitoring a disease over time and independent of appointment status. These automatic rules could be particularly useful for monitoring PROMs before and after certain interventions or treatments. For certain PROMs, integration with the EHR may not be possible because of copyright infringement, difficulty adapting certain questions to a digital platform, or lack of studies validating the electronic administration of the particular PROM.
Similar to distribution, EHRs should provide multiple ways for patients to complete PROMs. These methods make PROMs easily accessible to patients and help address differences in technological savvy, practice resources, and practice settings. This infrastructure can ensure the completion of PROMs by presenting them at various points of care. PROMs can be completed by patients at home or while admitted to a hospital via an EHR patient portal, at check-in on a tablet or kiosk, or in a clinic in a front-end software application. For PROMs available in an EHR patient portal, patients could complete the questions by using the web or an app. Completion of PROMs before the clinic visit is ideal for minimizing disruptions to clinic workflow and enhancing efficiency. Completion at check-in on a tablet or kiosk is ideal for PROMs that may be unanswered in the EHR patient portal or those with sensitive questions (eg, suicidality) that require in-person completion. If a tablet or kiosk is not available, PROMs can be answered on a clinic workstation directly in a front-end software application by either the patient or provider. Finally, hospitalized patients can complete PROMs using an EHR patient portal that could be geared toward the inpatient setting. For all modalities, PROM responses should be viewable by providers or ancillary staff in real-time and be configured for automatic scoring, further reducing inefficiency.
To allow for seamless integration into their preexisting workflow, providers should be able to review PROM responses in several formats. For example, PROM responses can be viewed in a flowsheet, encounter-level report, or electronic message basket. These visualization methods are typically encounter-based and are ideal for point-of-care review and clinical decision-making. Scores can also be viewed over time, which allows clinicians to easily compare a patient’s PROM scores with other relevant data, such as medications, laboratory results, diagnostic testing reports, procedures, or surgeries. Patient responses and/or scores need to be easily pulled into documentation and, if desired, communicated with other providers via a secure message or by creating a letter to be faxed or mailed.
The EHR infrastructure must facilitate the review, analysis, and quality control of population-level PROMs data. For example, operational reports need to provide robust data for PROMs, such as total scores, minimally important differences, and questionnaire completion rates. Easily accessible population-level PROMs data are necessary for a better understanding of the clinical outcomes, to facilitate patient-oriented research, and to inform public health initiatives.
Discussion
PROMs can be employed as screening tools in the general population and also as tools to assess disease severity and response to therapy. Incorporating these PROMs into the pediatric primary care setting could be especially impactful in children with respiratory diseases.
PROMs specific to asthma management are widely used by primary care providers and pulmonary subspecialists. Over the last 2 decades, several asthma-specific PROMs that target the pediatric population have been developed and validated. Despite the benefits PROMs offer in disease management, challenges persist. Limitations include (1) less-than-ideal correlations between many of the PROMs necessitating use of multiple instruments, (2) inconsistent relationships between PROMs and objective outcomes such as spirometry, (3) a lack of data evaluating the effect of disease severity, age, race, ethnicity, cultural beliefs, and health literacy, (4) inadequate translation of PROMs into other languages, which may contribute to health disparities, and (5) the potential for survey fatigue. Further research is also needed to define the optimal duration of time needed to demonstrate changes in the various PROMs because it may take longer to detect changes in PROMs than in physiologic measures.
Pediatric nephrologists have implemented logical, stepwise, and scientifically rigorous qualitative and quantitative approaches to incorporate patient, caregiver, and health care provider input in subspecialty- or condition-specific PROMs. Ongoing multicenter studies or national consortiums that study pediatric lung diseases could be leveraged to conduct research designed to develop, validate, and assess the utility of PROMs to assess childhood respiratory health.
Incorporating PROMs designed to assess pediatric respiratory health into the primary care setting is certainly possible and worthwhile. To achieve this, we will need to (1) focus on best practices for using a screening tool for children in the outpatient setting, (2) maintain attention to overcome diverse provider-, patient-, and health-care-system-level barriers to the adoption of screening tools, and (3) guard against the potential for health inequities in the implementation of screening tools. Operationalizing will require a planning process to determine the specifics of how screening tools will be done, when they will be done, who they will be used for, how they will be tracked, and how they will be shared with other health care providers for consultation (such as pediatric pulmonologists) if needed. Once available for use, a careful evaluation, review, and analysis of the adoption of PROMs by providers could shorten the timeline for translation to primary care.
In addition, the electronic health record must be harnessed for the successful adoption of PROMs for children with pediatric lung diseases. The process to use PROMs should be easily adoptable in all primary pediatric care settings regardless of the resources available. The availability of a variety of technologies can make PROMs more widely available to all children, their families, and other health care providers, thus eliminating the digital divide and reducing health disparities. Options include practice websites, patient portals, registration kiosks, telemedicine platforms, smartphone apps, school-issued devices, office-based tablets, and remote patient monitoring instruments. EHR tools are critical to the success of PROMs implementation. In particular, the EHR infrastructure should promote patient and clinician accessibility, efficiency, and flexibility to meet the varying needs of all stakeholders.
Conclusions
Over the last few decades, substantial progress has been made in defining, measuring, and promoting the use of PROMs in the pediatric population. Despite a long-standing history of utilizing PROMs for the evaluation and management of childhood asthma, pediatric pulmonologists still lag behind other pediatric subspecialists in the use of PROMs. Yet, arguably, many pediatric lung diseases could be better characterized by a patient’s symptoms, behaviors, and functional limitations than by physiologic measurements.
To achieve our goal of improving pediatric respiratory health at the community level, PROMs that survey key domains of childhood lung health could be implemented in the primary care setting. However, given the distinctive neurocognitive and pulmonary developmental issues at different ages, designing such PROMs and promoting their adoption by providers will not be an easy task. The need to involve the caregivers and to understand how these PROMs can supplement physiologic measures (such as lung function testing) will add other layers of complexity. The development of a successful conceptual framework is built on successes in general pediatrics and other pediatric subspecialties, instruments tailored to the child’s age group, and a focus on health equity and inclusion. Technology, such as the EHR, will be also critical for success. Ultimately, an integrative approach to systematically address numerous barriers at the level of the provider, patient, and health care system will be needed to attain this goal and achieve sustainability.
Drs Bacharier, Buckley, Amaral, and Sirota wrote the initial drafts of individual sections of the manuscript; Drs Rosas-Salazar and Moore conceptualized the manuscript, compiled and revised the individual sections, included additional relevant text, and created the final manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
The views expressed in this article are those of the authors and do not necessarily represent those of the National Institutes of Health or the US Department of Health and Human Services.
FUNDING: The virtual workshop on which this report is based was funded by the US National Heart, Lung, and Blood Institute (NHLBI). The NHLBI approved the workshop concept and provided administrative support for the workshop.
CONFLICT OF INTEREST DISCLOSURES: Dr Bacharier reports personal fees from GlaxoSmithKline, Genentech/Novartis, DBV Technologies, Teva, Boehringer Ingelheim, AstraZeneca, WebMD/Medscape, Sanofi/Regeneron, Vectura, Circassia, Kinaset, Vertex, and OMPharma and royalties from Elsevier outside the submitted work. The other authors have indicated they have no potential conflicts of interest relevant to this article to disclose.
Comments