Identifying Cases of Undiagnosed, Clinically Significant COPD in Primary Care Qualitative Insight from Patients in the Target Population

SUPPLEMENTARY INFORMATIONIdentifying Cases of Undiagnosed, Clinically Significant COPD in Primary Care: Qualitative Insight from Patients in the Target Population NK Leidy, PhD1; K Kim,

SUPPLEMENTARY INFORMATION

Identifying Cases of Undiagnosed, Clinically Significant COPD in Primary Care:

Qualitative Insight from Patients in the Target Population

NK Leidy, PhD1; K Kim, MPH1; EJ Dansie, PhD1; BP Yawn, MD, MSc2; DM Mannino, MD3;

BM Thomashow, MD4; RG Barr, MD4; SI Rennard, MD5; JF Houfek, PhD5;

MK Han, MD, MS6; CA Meldrum, PhD6; BJ Make, MD7; RP Bowler, MD7; AW Steenrod, MPH1;

LT Murray, MPH1; JW Walsh8; F Martinez, MD, MS9; for the High-Risk-COPD Screening Study Group10

1Evidera, Bethesda MD; 2Olmsted Medical Center, Rochester MN; 3University of Kentucky, Lexington KY; 4Columbia University, New York NY; 5University of Nebraska, Omaha NE;

6University of Michigan, Ann Arbor MI; 7National Jewish Health, Denver CO; 8COPD Foundation, Washington DC; 9Weill Cornell Medical Center, New York, NY

10High-Risk-COPD Screening Study Group:

Rebecca Copeland, BS, University of Kentucky; Tim Dorius, MD, University of Nebraska Medical Center; David Hengerer, BA, Evidera; Patricia Jellen, RN, MSN, New York

Presbyterian Hospital; Marge Kurland, RN, Olmsted Medical Center; Karen Malley, BA, Evidera; Jason Shiffermiller, MD, MPH, University of Nebraska Medical Center; Christina Schnell, BA, CCRC, National Jewish Health; Lori Silveira, PhD, MS, National Jewish Health; Sonja Stringer, MPH, Evidera; Deb Sumnick, PBT, University of Nebraska; Randel Plant, COPD Foundation; Jennifer Underwood, CCRP, National Jewish Health; Beth Whippo, MSN,

RN-BC, New York Presbyterian Hospital

Cognitive Interviews

This supplement describes the methods and results of the cognitive interview phase of instrument development The purposes of these interviews are to assess respondent comprehension of the draft instructions and pool of candidate questions for the screener and make adjustments for added clarity before the items are administered to a large sample for quantitative testing and item-reduction

Study Design

One-on-one in-person cognitive interviews were conducted with adults with and without COPD Data on value consistency and ease of use of electronic and manual PEF devices were also collected

Sample

The sample included individuals from the target population, with and without experiences with the constructs assessed in the draft item pool Three sets of individuals were recruited Set 1 included those with a recent COPD diagnosis (prior 36 months) and an FEV1/FVC Ratio <0.70, with or without the initiation of maintenance therapy These individuals should be able to

evaluate the clarity of the items from their perspective as a person with a recent diagnosis of, and familiarity with, COPD Participants in Set 2 did not have COPD, but were pre-screened for risk factors and included both smokers (current or former, ≥10 pack-years) and non-smokers (≤100 cigarettes/pipes/cigars in a lifetime) Consistent with the elicitation-phase screening, risk factors included the following: shortness of breath with activity; cough; phlegm (sputum) in the absence

of a cold; wheezing; history of asthma; exposure to second-hand tobacco or other kinds of smoke

at home or work; exposure to dust, gases, or dirty air at work; colds generally settling in the chest; one or more respiratory events that resulted in missed work or school and may have led to

a clinic visit or emergency room visit; admittance to the hospital with breathing problems; and serious childhood breathing conditions This set of participants would be able to speak to the clarity of items associated with various risk factors, but would be nạve to the personal diagnosis and manifestations of COPD Finally, participants in Set 3 included smokers and never smokers

without COPD With the exception of smoking, this group was not pre-screened, to more closely represent the experience of individuals asked to complete a screening tool for COPD without prior reference or exposure to specific risk factors

Additional inclusion criteria were as follows: ≥40 years or older; willing and able to participate

in the cognitive interview and perform peak flow testing; able to read and speak in English; and willing and able to provide written informed consent Individuals were not eligible if they had participated in the elicitation phase of the project, had been hospitalised for a respiratory

infection the prior 30 days, or, in the judgment of the clinician or clinical coordinator, had cognitive impairment, hearing or sight difficulty, or severe psychopathology that could interfere with ability to participate in the study

Procedures

The protocol was approved by a central IRB as well as the local IRB at each participating site COPD and non-COPD participants were recruited from two pulmonary sites and one primary care clinic; non-COPD participants were also recruited from the general population, through an advertisement in a local newspaper All data were gathered in the United States Convenience sampling was used, with clinical members of the Study Group approaching potential participants

by phone or during clinic visits.Efforts were made to recruit a diverse sample with regards to age, smoking status and history, gender, race/ethnicity, and educational level Prior to data collection, written informed consent was obtained for all participants

Experienced and trained research staff used a semi-structured interview guide to conduct these 1:1 interviews, conducted in a private room in the clinic These cognitive interviews were

designed to gather data on the clarity of the instructions and candidate items, including the extent

to which each item was understood and interpreted as intended Participants were instructed to read the instructions aloud and explain them to the interviewer, in their own words If the

instructions were unclear or the interpretation was not consistent with the intent, alternative words or phrases were discussed to inform future editing Participants then read each item aloud, recorded a response, and briefly explained what the item asked and how they selected their response If the item was not clear to the participant or their interpretation was not consistent with the intent, alternative words or phrases were discussed to inform future editing If previous interviews resulted in an edited or new item, this information was presented to the participant on

a flashcard, to determine if the revisions were clear and interpreted correctly Interviews lasted 30–75 minutes; field notes were taken during or following the interview to supplement transcript data as needed All interviews were audio-recorded and transcribed verbatim, with the data cleaned and de-identified for analyses

Following the interview, PEF data were gathered from each participant using both an electronic (Vitalograph® Asma-1 USB) and manual (Vitalograph® AsmaPlan mech PFM) PEF device in random order; 3 acceptable readings of each were recorded SafeTway disposable mouthpieces were provided and changed between participants Both participants and staff administering the PEF tests rated each device in terms of ease of use for the participant (very easy, easy, neither difficult nor easy, difficult, or very difficult to use)

To characterise the sample, participants also completed a brief sociodemographic form, the generic SF-361 and the CAT2 which generally took less than 10 minutes to complete

Data Analysis

Content analyses were used to analyse the data, with a content grid used to capture item-by-item comprehension across participants An item was considered understood if the participant was able to read, respond, and provide a logical explanation for their answer Specific comments and suggestions were also tracked All modifications to the draft instrument and reasons for those modifications were recorded in an item tracking matrix

Descriptive statistics were used to summarise sample demographic and clinical characteristics, PEF values, and PEF ease of use Due to the small sample sizes, statistical testing was not performed on the PEF data

RESULTS

The sociodemographic and clinical characteristics for the study sample are shown in Tables 1 and 2, respectively Most participants demonstrated a clear understanding of the instrument instructions One change was made for added clarity: an explicit phrase was added telling participants how to indicate their answer for each question (“by marking an X in the box

provided”)

Most of the candidate items were easily understood Participants offered insight into how

specific terms or risk factors were interpreted, leading to modifications in the candidate items for greater clarity For example, when asked about exposure to smoke, several defined “smoke” and

“second-hand smoke” differently, interpreting the former in terms of the environment (fireplace, work, cooking) and the latter in terms of cigarette smoke, specifically A question about

shortness of breath with moderate to intense activity used gardening as one example; several subjects suggested changing the term to “shoveling dirt or snow”, to clarify and generalise the task When asked about hospitalisations, emergency room or clinic visits for a “lung problem”,

several participants indicated this would exclude visits for shortness of breath because shortness

of breath was not viewed as (associated with) a lung problem Finally, several participants, both COPD and non-COPD, suggested adding a question related to waking up at night feeling short of breath

Electronic and manual PEF devices were rated easy or very easy to use by most participants (93%; 96%) and test administrator (85%; 96%) PEF values are shown in Table 3, with the electronic device yielding descriptively higher values than the manual device

CONCLUSION

This work was part of a multi-method approach for developing a screening method for

identifying cases of undiagnosed, clinically significant COPD in primary care The cognitive interviews described here were designed to evaluate the clarity of candidate items in a sample of individuals with diverse experiences with COPD and risk factors for the disease Consistent with the elicitation sample, risk factors were similar across the three sets of subjects, including the set that was not pre-screened Participants provided insight into how the items were interpreted, and their responses were used to modify candidate items for greater clarity The candidate items are ready for quantitative testing and item-reduction in a large sample of cases and controls

1 Ware JE, Jr., Sherbourne CD The MOS 36-item short-form health survey (SF-36) I

Conceptual framework and item selection Med Care 1992; 30: 473-483.

2 Jones PW, Harding G, Berry P, Wiklund I, Chen WH, Kline Leidy N Development and

first validation of the COPD Assessment Test Eur Respir J 2009; 34: 648-654.

Table 1 Sample Demographic Characteristics

(N=27)

Recent COPD (Set 1) (N=8)

No COPD

>2 Risk Factors (Set 2) (N=9)

No COPD

No Pre-Screen (Set 3) (N=10)

Ethnicity, n (%)

Marital Status, n (%)

Other (single, divorced,

Other (unemployed, homemaker/

Education Level, n (%)

1 Categories are not mutually exclusive

Table 2 Sample Clinical Characteristics

(N=27)

Recent COPD (Set 1) (N=8)

No COPD

> 2 Risk Factors (Set 2) (N=9)

No COPD

No Pre-Screen (Set 3) (N=10) Smoking Status and History

Age started smoking, mean

Other Risk Factors, n (%)

Family history of breathing

(%)

Breathing Events Past Year, n (%)

Breathing Event Impact, n (%)

Characteristic Total

(N=27)

Recent COPD (Set 1) (N=8)

No COPD

> 2 Risk Factors (Set 2) (N=9)

No COPD

No Pre-Screen (Set 3) (N=10)

Emergency room visit/urgent

Non-Respiratory Health

Conditions, n (%)

Health Status, mean (SD), median

SF-36 – Physical Component

SF-36 – Mental Component

COPD Assessment Test (CAT) 4 15 (8); 15 21 (8); 22 15 (6); 15 11 (7); 11

1 Categories are not mutually exclusive

2 Chest congestion, discomfort, tightness, pain, or wheeze

3 Scale: 0–100; higher scores are better

4 Scale: 0–100; higher scores are worse

Table 3 Mean (SD) Peak Flow by Device Type and Set

Recent COPD (Set 1) (N=8)

No COPD

> 2 Risk Factors (Set 2) (N=9)

No COPD

No Pre-Screen (Set 3) (N=10)

1 Paired; no statistical tests were performed due to small sample sizes