Interpreting respiratory oscillometry in adults with asthma or COPD: findings of an international Delphi study

Abstract

Background

Respiratory oscillometry measures the physiological effort and mechanics of moving air in and out of the lungs during normal breathing. It provides complementary information to spirometry. Uncertainty regarding the interpretation of oscillometry is a barrier to routine use. The aim of this study was to aid in oscillometry interpretation among adults with asthma or COPD by generating expert consensus statements.

Methods

A Delphi method was used to develop consensus statements regarding the clinical use of oscillometry in adults to identify abnormal lung function, bronchodilator response and minimal clinically important differences. Initial statements were refined in the brainstorming round. Statements were assessed by 60 pulmonologists over three rounds, with consensus defined as ≥70% agreement.

Results

Pulmonologists agreed that oscillometry is clinically useful to assess abnormal lung function and its severity, and to measure bronchodilator response. High consensus was reached for resistance at 5 Hz (R₅, 85%), reactance at 5 Hz (X₅, 79%) and area under the reactance curve (A_X, 77%) based on z-scores, where >1.64 was considered abnormal for R₅ and A_X and < −1.64 was considered abnormal for X₅. For measuring bronchodilator response, good agreement was based on using percentage change for R₅, X₅ and A_X.

Discussion

This international Delphi study combined evidence-based and expert opinion to inform clinicians in the interpretation of respiratory oscillometry. Focusing on a few key parameters of oscillometry will allow clinicians to become confident in its everyday use to assess abnormal lung function, grade severity of impairment, monitor progression over time and assess bronchodilator response.

Shareable abstract

This international Delphi study informs clinicians in the use of respiratory oscillometry; focusing on a few key parameters will build confidence in its everyday use to assess abnormality and severity, monitor changes, and determine bronchodilator response https://bit.ly/3HPoPgI

Introduction

Respiratory oscillometry (also known as the forced oscillation technique) is a noninvasive, effort-independent lung function test that is sensitive to all airways, including the peripheral airways [1]. Oscillometry measures respiratory system impedance by superimposing oscillatory pressure waves at the mouth during tidal breathing. Oscillometry assesses how difficult it is to move air in and out of the lungs during normal breathing. The key measures are the resistance of the airways and tissues to flow (R_rs), a measure of airway calibre, and the respiratory system reactance (X_rs), which is a measure of stiffness of the respiratory system, including lung parenchyma and the chest wall [1, 2]. Hence, X_rs is related to the degree of gas trapping and hyperinflation, ventilatory inhomogeneities in the lungs of obstructive lung disease [3, 4], and the reduced lung compliance of fibrotic lung disease.

As a more sensitive indicator of small airways dysfunction, oscillometry complements spirometry by providing clinicians with additional information about the patient's airway physiology, assisting in the diagnosis of asthma in patients with preserved spirometry [5, 6], and identifying abnormality among smokers with preserved spirometry [1]. The addition of oscillometry to forced expiratory volume in 1 s (FEV₁) improves the identification of poor asthma control and elevated exacerbation risk compared with spirometry alone [7–9]. In the ATLANTIS study, the frequency dependence of resistance (R_5–20), area under the reactance curve (A_X) and reactance at 5 Hz (X₅) were identified as indicators of small airways dysfunction and potential indicators of poor asthma control and exacerbations [8, 10]. For patients with COPD, X_rs is related to the degree of gas trapping, hyperinflation and ventilatory inhomogeneities in the lungs, and reflects the amount of communicating lung volume as measured by plethysmography [3] and functional imaging [11, 12]. Reactance measures have been related to symptoms [13, 14], exacerbation risk [14] and mortality [15, 16].

Routine clinical use of oscillometry in adults with asthma or COPD is growing with the wider availability of devices and the publication of technical standards [17] and normative population data [18–21]. Despite this, barriers to broader adoption exist, which include uncertainty regarding the interpretation and meaning of the numerical outputs for oscillometric parameters, in part due to the lack of evidence-based cut-offs and minimal clinically important differences (MCIDs) [21, 22].

The oscillometric-derived parameters frequently reported in adults include the resistance (R₅) and reactance (X₅) at low frequencies (5 Hz), the frequency dependence of resistance, i.e. the resistance reduction between 5 and 19 or 20 Hz (R_5–19 or R_5–20), and A_X (table 1) [1]. Within-breath [23] and intra-breath [24] analysis of resistance and reactance are also increasingly used.

TABLE 1.

Overview of key oscillometry parameters and clinical implications in airways disease [1, 10]

Oscillometry parameter	What it measures	What an abnormal finding may indicate
R 5	Changes in airway resistance or calibre	Airway narrowing, obstruction
R5–20/R5–19	Airway heterogeneity, peripheral airway resistance, potentially confounded by upper airway shunting or obesity	Possible small airways dysfunction
X 5	Stiffnesses of the respiratory system (lung and chest wall) during tidal breathing (dependent on the accessible lung volume during the measurement or parenchymal compliance)	Small airways dysfunction, possibly accompanied by ventilation heterogeneity and/or lung derecruitment (airway closure)
AX	Summative measure of the respiratory system stiffness across a range of frequencies (similar to X5)	Similar to X5

R₅: resistance at 5 Hz; R_5–20/R_5–19: frequency dependence of resistance; X₅: reactance at 5 Hz; A_X: area under the reactance curve.

In the absence of definitive data to develop evidence-based guidelines on the interpretation of oscillometry, a Delphi study was conducted among clinicians who routinely use the technique. The aim of this study was to aid in the interpretation of respiratory oscillometry in clinical practice among adult patients with asthma or COPD by generating high-quality expert consensus statements on which oscillometric parameters are routinely used and what their associated cut-offs are.

Methods

An international panel of experts participated in a three-round Delphi study to determine consensus on the interpretation of oscillometry in adults with asthma or COPD. This study was approved by Bellberry Human Research Ethics Committee (2023-11-1410). All participants provided written informed consent. The study was registered on www.clinicaltrials.gov (NCT06313372). Surveys were administered anonymously using SurveyLet (Calibrum Inc.), an online survey platform that uses the Delphi study technique.

Participant recruitment

Participants were pulmonologists with expertise in the use of oscillometry in clinical practice. Inclusion criteria were pulmonologists in active clinical practice for ≥12 months who routinely use oscillometry in the management of adults with asthma or COPD, defined as an average of at least two tests per week. Pulmonologists were excluded if they only used oscillometry for research or were an employee or stockholder of a pharmaceutical company, oscillometry manufacturer or tobacco/vaping company.

The study steering committee (L.P.C., B.T. and C.S.F.) identified clinicians from those who had published in the literature on oscillometry using a snowballing recruitment technique, such that invited clinicians could nominate other pulmonologists to participate. A smaller group of experts were invited to participate in the brainstorming round to develop the consensus survey.

Literature search and brainstorming

A systematic literature review was conducted to identify published papers on the use of oscillometry in adults with asthma or COPD that provided information about cut-off values, reference equations or MCIDs. Searches were carried out using EMBASE and MEDLINE, covering inception to 19 December 2023. Search terms were respiratory oscillometry AND reference equations OR interpretation OR cut-offs OR minimal clinically important difference.

Published data were reviewed by the steering committee and used to draft the brainstorming survey. This brainstorming survey was circulated to the selected experts for review and refinement.

Rounds 1 to 3 and determination of consensus

To obtain consensus on the interpretation of oscillometry, three iterative survey rounds were undertaken. In each round, an online survey comprising statements and questions regarding oscillometry was sent to all participants. In round 1, the survey included participant demographic questions. The survey predominantly consisted of statements that were assessed for the level of agreement using a 6-point Likert scale (agree strongly, agree moderately, agree slightly, disagree slightly, disagree moderately and disagree strongly). Participants could provide comments on each question, which the steering committee reviewed to determine if the survey required modification before the next round. Questions were repeated in subsequent rounds if consensus was not achieved. In rounds 2 and 3, participants could review the aggregated results from the previous round.

The consensus threshold was predefined as ≥70% of the participants giving the same response for that question. For questions that used the 6-point Likert scale, consensus was assessed by summing the results as follows:

• Agreement was defined as responses of “agree strongly” plus “agree moderately”.

• Neutral was defined as “agree slightly” plus “disagree slightly”.

• Disagreement was defined as “disagree strongly” plus “disagree moderately”.

Statistical analysis

Statistical analysis was performed using the SurveyLet statistical analysis software. The distribution of responses for each question was produced following each Delphi round.

The level of agreement was assessed using the consensus score for Likert scale questions. Consensus scores (centrality statistics) for quantitative variables with normal distribution were reported as mean±sd. Other quantitative variables were reported as medians and interquartile ranges. Group stability, indicating the strength of consensus, was calculated using percentages for categorical questions. Descriptive statistics were used to analyse demographic data.

Results

Literature review

The systematic literature review identified 210 published papers with potentially relevant information on interpreting oscillometry. After removing duplicates, 188 papers were reviewed and 59 papers contained relevant information (supplementary table 1).

Participants

Of the 12 experts invited to participate from the brainstorming round, 10 participated, with one expert excluded and one not commencing the survey. Of the 197 clinicians invited to participate, 67 were enrolled, three withdrew and four were excluded. Of the 60 study participants, 55 (91.7%) completed all three survey rounds (figure 1).

FIGURE 1.

Study flow diagram and survey completion rates. Note: one participant that completed the brainstorming survey did not complete the round 1 survey but was not excluded from the study

.

The clinicians in this study had extensive experience in using oscillometry: 66% had used oscillometry for ≥5 years, with a median of 10 tests per week (table 2).

TABLE 2.

Demographic data of study participants (n=59)

Age, years
Mean±sd	49.3±12.5
Range	30–66
Gender #
Male	38 (63.3%)
Female	22 (36.7%)
Geographic region#
Asia (India, Japan, Korea, Malaysia and Taiwan)	7 (11.7%)
Australia and Oceania (Australia and New Zealand)	9 (15.0%)
Europe (Austria, Bosnia, Denmark, France, Greece, Italy, Russia, Serbia, Sweden and UK)	32 (53.3%)
North America (Canada, Mexico and USA)	9 (15.0%)
South America (Ecuador)	1 (1.7%)
Middle East and Greater Arabia (UAE)	2 (3.3%)
Years practicing as a respiratory specialist/pulmonologist
1 to <5 years	6 (10.2%)
5 to <10 years	10 (16.9%)
>10 years	43 (72.9%)
Years using respiratory oscillometry in clinical practice
0 to <1 years	5 (8.5%)
1 to <5 years	15 (25.4%)
5 to <10 years	22 (37.3%)
>10 years	17 (28.8%)
Practice environment
Tertiary or university hospital	50 (84.7%)
Regional, district hospital	5 (8.5%)
Private practice	15 (25.4%)
Other	2 (3.4%)
Clinical setting in which oscillometry is routinely used
Outpatient	33 (55.9%)
Inpatient	2 (3.4%)
Both	24 (40.7%)
Approximate percentage of work spent caring for adult patients with asthma or COPD
Mean±sd	56.3±21.9%
Range	10–90%
Average respiratory oscillometry tests performed per week
Median (IQR)	10 (11)
Type of respiratory oscillometry test used in clinical practice
Impulse oscillometry	43 (72.9%)
Pseudo-random (forced oscillation technique)	21 (35.6%)
Authored a publication(s) on respiratory oscillometry in the past 5 years#
Yes	34 (56.7%)
No	26 (43.3%)

Data are presented as n (%), unless otherwise stated. UAE: United Arab Emirates; IQR: interquartile range. ^#n=60 (one participant that completed the brainstorming survey did not complete the round 1 survey but was not excluded from the study).

Consensus

Table 3 summarises the consensus clinical guidance. Figure 2 shows the level of agreement achieved for each statement. Of the 18 statements that achieved consensus, 12 were achieved in round 1. Supplementary table 2 summarises how consensus evolved over the study.

TABLE 3.

Summary of guidance on the interpretation of respiratory oscillometry

Percentage agreement	Recommendation
Defining abnormal lung function
85%	Use z-scores to identify an abnormal R5
85%	The cut-off for an abnormal R5 is a z-score >1.64
79%	Use z-scores to identify an abnormal X5
92%	The cut-off for an abnormal X5 is a z-score < −1.64
77%	Use z-scores to identify an abnormal AX
79%	The cut-off for an abnormal AX is a z-score >1.64
78%	Use % predicted to identify an abnormal R5
70%	The cut-off for an abnormal R5 is a % predicted >150%
81%	Use z-scores to grade the severity of abnormal lung function: Mild is a z-score >1.64 and ≤2.5 Moderate is a z-score >2.5 and ≤4 Severe is a z-score >4
Defining a bronchodilator response
72%	The cut-off for a significant bronchodilator response for R5 is a change in z-score >1.4
84%	Use % change in R5 to assess a significant bronchodilator response
79%	Use % change in X5 to assess a significant bronchodilator response
91%	Use % change in AX to assess a significant bronchodilator response
Defining an MCID
71%	Use z-scores for X5 to assess an MCID between visits
77%	Use % change in AX to assess an MCID between visits

R₅: resistance at 5 Hz; X₅: reactance at 5 Hz; A_X: area under the reactance curve; MCID: minimal clinically important difference.

FIGURE 2.

a) Level of consensus for defining abnormality. b) Level of consensus for defining a bronchodilator response (BDR). c) Level of consensus for defining a clinically meaningful change between visits. A_X: area under the reactance curve; R₅: resistance at 5 Hz; R_5–20: frequency dependence of resistance; X₅: reactance at 5 Hz.

Clinical utility of oscillometry

It was agreed that oscillometry is clinically useful for the following:

• 1) grading the severity of lung function impairment (81%)

• 2) measuring bronchodilator response (BDR) (88%)

• 3) assessing clinically meaningful changes in lung function between visits (79%)

Key oscillometry parameters

Participants ranked the order of clinical importance for seven commonly used oscillometric parameters. The four most important parameters identified were R₅, R_5–20, X₅ and A_X.

The use of these parameters and their associated cut-offs overall did not differ when conducting oscillometry among adults with asthma or COPD.

Reference equations

The most used reference equations were derived from Oostveen et al. [18], used by 61.7% of clinicians. Other common sources of reference equations were Jetmalani et al. [19], Schulz et al. [21] and Berger et al. [20], each used by ∼25% of clinicians.

Defining abnormal lung function

Measures of resistance (R₅ and R_5–19 or R_5–20) and reactance (X₅ and A_X) are commonly used to assess abnormal lung function (supplementary table 3). Agreement was achieved for the use of z-scores to determine abnormality for R₅, X₅ and A_X as well as the corresponding z-score cut-offs (table 3 and figure 2a).

To assess abnormal R₅, % predicted was used by 56.7% of respondents, with a cut-off of >150% defining abnormality.

There was no consensus to support the use of absolute values to define abnormal lung function for any parameter.

Measuring BDR

Pulmonologists agreed (88%) that measuring BDR with oscillometry is clinically useful. However, defining a significant BDR with oscillometry did not reach consensus, with approximately equal proportions of pulmonologists recommending a change in either “resistance or reactance” (47%) or in both “resistance and reactance” (45%).

BDR was mostly assessed by measuring a change in R₅ (98.3%), followed by X₅ (80%) and A_X (70%). Cut-offs for defining a significant BDR did not reach consensus, with the most common cut-offs being a percentage change >40% for R₅, >50% for X₅ and >80% for A_X (49%, 49% and 44% agreement, respectively).

Using z-scores to define BDR did not achieve consensus for R₅ or X₅; however, there was minimal disagreement with this approach (<10%). Despite this, it was agreed that a change in the z-score for X₅ of >1.4 indicated a significant BDR. The cut-off for R₅ of >1.4 had 65% agreement and minimal disagreement (<5%) (figure 2b).

There was no consensus support for the use of absolute values for assessing BDR.

Defining an MCID between visits

Pulmonologists agreed (79%) that a change in oscillometry results between visits beyond day-to-day variation is clinically useful. This was mostly assessed by tracking R₅ (72%) and R_5–20 (65%), and less commonly by X₅ (58%) and A_X (54%).

For R₅, there were equal levels of support for using z-scores (68%) or percentage change (68%) to identify an MCID, but these parameters and their corresponding cut-offs did not achieve consensus. For R_5–20, the use of percentage change achieved some agreement (57%) but not consensus (figure 2c).

Agreement was achieved for the use of percentage change to identify an MCID for A_X (77%) and to use z-scores for X₅ (71%); however, the corresponding cut-offs did not achieve consensus. Participants often based MCID cut-offs on the values used to identify a BDR, but most recommended cut-offs were lower than those used for BDR (supplementary table 4).

There was no consensus support for the use of absolute values to identify an MCID for any parameter.

Discussion

This is the first Delphi study to provide expert guidance on the use of oscillometry in adults with asthma or COPD. Oscillometry can be routinely used in conjunction with other lung function tests to provide more information about the patient's lung function, including insights into small airways dysfunction [10]. As with other lung function tests, clinicians should routinely use a small number of parameters to guide their interpretation [25], with experts supporting the routine use of R₅, X₅ and A_X (figure 3).

FIGURE 3.

Clinical guidance for respiratory oscillometry. Bronchodilator response (BDR) was assessed based on percentage change versus baseline (prebronchodilator). The most commonly used BDR cut-offs were those published in the technical standards [17] but all were below the consensus threshold. A_X: area under the reactance curve; R₅: resistance at 5 Hz; X₅: reactance at 5 Hz.

Defining abnormality

Oscillometry can be used to identify abnormal lung function, with R₅ indicating the presence of airway obstruction and reactance (X₅ and A_X) indicating the presence of small airways dysfunction. The identification of small airways dysfunction using abnormal X₅ and/or A_X results is clinically relevant because these abnormalities help to identify patients at increased risk of poor asthma control and exacerbations [7, 9, 26]. In addition, small airways dysfunction is associated with more severe bronchial hyperresponsiveness [27], nocturnal asthma [28], exercise-induced asthma [28, 29] and reduced quality of life [7, 10]. For those clinicians that adopt a treatable traits approach, the presence of small airways dysfunction could be treated more effectively using extra-fine particle inhalers, which provide a more uniform medication deposition along the bronchial tree, including the small airways [26, 30, 31].

As with other lung function tests, z-scores should be used as the primary means to interpret results [17, 25]. For R₅ and A_X, abnormality cut-offs were set at a z-score of >1.64, representing a 5% chance of a false-positive finding in healthy individuals [25]. The z-score cut-off for an abnormal X₅ is < −1.64 because X₅ values are negative, and a lower (more negative) value indicates worse lung function.

Although some clinicians used % predicted to measure an abnormal R₅, z-scores are preferred because % predicted does not account for other determinants of lung function such as age, sex and height [25]. Using % predicted to define abnormality for X₅ is discouraged because X₅ values are very small and often close to zero, which can cause the % predicted values to become very large and disproportionate to the true physiological difference.

Assessing the severity of lung function impairment can help guide management, and z-scores were endorsed for this grading, with cut-offs that define mild, moderate and severe impairment being the same as those applied in spirometry [25]. This recommendation is not evidence based but is statistically sound. A single registry-based cohort has reported on such cut-offs for oscillometry in health and a variety of lung conditions [32]. The accuracy and validity of a z-score depends on whether the reference equation is applicable to the patient population being studied. Furthermore, z-score cut-offs for defining abnormality and the severity of impairment in oscillometry are not yet anchored to clinical outcomes, such as exacerbation or hospitalisation risk, and require further longitudinal validation studies [33].

When designing the Delphi survey, there were no widely adopted reference equations for R_5–19 or R_5–20; therefore, z-scores are not generally used when interpreting this parameter. Most clinicians use an absolute value to define abnormality. Round 1 determined the cut-offs used and, in round 2, consensus was assessed for the median cut-off of 0.1 kPa·s·L⁻¹. This was revised to 0.07 kPa·s·L⁻¹ in round 3, based on feedback that this value is most commonly used in several studies [28, 34, 35]. This change between rounds may have contributed to the lack of consensus.

BDR

Oscillometry can be used to assess BDR in adults as it can identify bronchodilation in more patients than spirometry [7, 18]. Bronchodilation, as identified by changes in reactance (X₅ and A_X), measures a different aspect of lung function to spirometry, reflecting the opening up of smaller airways and potentially improved ventilation inhomogeneity during tidal breathing. BDR, as assessed by X₅ and/or A_X, is clinically relevant because it is correlated with asthma control and may identify more patients with poor asthma control than spirometry [7].

Consensus was not reached on whether BDR should be defined as a change in both resistance and reactance or as a change in only one parameter. Given that spirometry defines BDR as a change in either FEV₁ or forced vital capacity [25], it is reasonable to apply the same approach to oscillometry until definitive evidence emerges.

One of the challenges in defining BDR is the limited normative data in health. The use of z-scores should be adopted as the primary methodology when more robust reference equations become available [17]. It is also recommended that the approach be anchored to clinical outcomes [25]. In the interim, defining BDR using the percentage change for R₅, X₅ and A_X is recommended. Although agreement for the cut-offs was not achieved, the most commonly recommended values were those proposed in the oscillometry technical standards [17]. One factor that may have contributed to the failure to achieve consensus on these cut-offs is the publication of new evidence since the technical standards supporting different cut-offs in adults were established [19, 36]. Until new evidence emerges, clinicians may choose to adopt the published cut-offs in the technical standards (figure 3) [17] or the proposed lower cut-offs of percentage change ≥32% for R₅ and ≥44% for X₅ [37]. Regardless of the chosen cut-off, its application in interpretating oscillometry should be consistent within each laboratory.

Tracking change between visits (MCID)

To date, there is only one study that provides evidence on MCIDs for oscillometry in adults with asthma [22]. Patients who had changes in R_5–20 and A_X greater than the MCID had greater improvements in patient-reported outcomes, and both were better predictors of symptom control and quality of life than changes in FEV₁. This reinforces the importance of assessing small airways dysfunction, especially when there is discordance between asthma control and spirometry.

The use of oscillometry to monitor patient progress over time is recommended. Clinicians generally approached this by monitoring the percentage change in oscillometry parameters and applying cut-offs, often based on those used to define BDR. These values are not necessarily interchangeable and further research to define MCIDs is needed. Such cut-offs will have to relate to relevant patient outcomes and not only the properties of a particular measurement, such as the coefficient of repeatability.

This is the first study to report the collective experience of clinicians using respiratory oscillometry to guide interpretation of this test in adults. Twice as many clinicians used impulse oscillometry than forced oscillation technique devices; however, subanalysis among impulse oscillometry users indicated that consensus outcomes were not influenced by device use. This is a reassuring finding with regard to the clinicians’ approach in interpreting oscillometry, acknowledging that known device-specific differences may limit direct comparison of individual values across devices. It is also noteworthy that previously published cut-offs for BDR are remarkably similar, irrespective of the type of device used [38]. Consensus was defined as ≥70%, a threshold used in other Delphi studies in asthma [39, 40] and to define core outcome sets [41].

A potential limitation is that the 70% consensus threshold may be considered low, with a systematic review of Delphi studies indicating that the median consensus threshold was ≥75% agreement [42]. Applying this higher threshold would not have altered our findings because all but three statements achieved consensus with >75% agreement.

The strengths of this study include the high proportion of participants that completed all three survey rounds and the geographic diversity of participants. The study was designed to assess the broad application of oscillometry in clinical practice and investigated the use of a range of cut-offs, including those that are problematic (e.g. absolute values and % predicted), as they do not account for the influence of age, sex and height on lung function [25]. The observation that 35.3% of statements achieved positive consensus reflects that most clinicians use the more physiologically meaningful and statistically robust measure, z-scores, to guide interpretation, and do not endorse those with inherent issues or an inadequate evidence base. The general principles of interpretation of z-scores in oscillometry should follow that for other lung function measurements, which include consideration for variability across devices, protocol and reference values, and disease states.

Research priorities to advance the clinical interpretation of oscillometry in adults with asthma or COPD include improved reference equations, as per the Global Lung Function Initiative [43], for both defining abnormal physiology and BDR. This along with future revision of oscillometry guidelines should address the question of whether there are valid thresholds that are useful for diagnosis of different airways disease phenotypes. There is a need to establish longitudinal MCIDs for key parameters and to anchor the grading of severity of impairment to clinical outcomes.

Broader adoption of oscillometry in clinical practice could be enhanced by incorporating the findings of this study into educational resources, such as clinical algorithms.

Conclusion

This international Delphi study directs clinicians to use respiratory oscillometry to assess abnormal lung function (small airways dysfunction), to grade the severity of impairment, to monitor progression over time and to assess BDR. Clinicians should interpret abnormal oscillometry based on z-scores for resistance or reactance and to interpret BDR using percentage change.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material

00398-2025.SUPPLEMENT2

Data availability

All data supporting the findings of this study are available within the paper and its supplementary material. Additional data, such as deidentified individual participant responses and study protocol, are available from the corresponding author following receipt of a methodologically sound proposal, to achieve the aims within the proposal, beginning 6 months and ending 24 months following article publication.