A Qualitative Analysis of Lung Cancer Screening Practices by Primary Care Physicians

Susan Henderson; Amy DeGroff; Thomas B. Richards; Julia Kish-Doto; Cindy Soloe; Christina Heminger; Elizabeth Rohan


J Community Health. 2011;36(6):949-956. 

In This Article


During May and June of 2009, prior to the release of the NLST initial findings, we conducted five telephone focus groups with 28 primary care physicians practicing in 15 states. Each group consisted of five or six participants, and the sessions lasted 75 min. A professional moderator with extensive experience in conducting telephone focus groups with physicians facilitated the discussions. We chose to conduct focus groups via telephone to increase participation rates and allow physicians from a wide variety of practice settings and geographic locations to participate. The study was supported by the Centers for Disease Control and Prevention (CDC). RTI International collaborated with CDC in conducting the study and coordinated the logistics of the focus groups.


Participating physicians who had specialized in family medicine or internal medicine or were considered a general practitioner, were licensed in the United States, practiced at least 20 h a week in a non-academic or non-government facility, and worked in cities with CT resources available were included. A professional focus group firm (Schlesinger Associates) recruited and screened physicians for participation using a national database it maintains. Prior to being assigned to a focus group, participants were asked, "Would you order lung cancer screening for an otherwise healthy, 50-year-old current smoker who has smoked one pack of cigarettes per day for the past 20 years?" Seventeen participants who responded "yes" were assigned to one of three "screening" groups, and 11 who responded "no" were assigned to one of two "non-screening" groups.

Data Collection and Focus Group Protocol

The moderator followed a semi-structured discussion guide to explore the following: (1) physicians' use of and attitudes toward cancer screening guidelines, as well as their attitudes toward the organization that issued the guidelines, (2) physicians' definition of "patient risk" in relation to lung cancer screening, (3) physicians' experiences with and responses to patients' requests for lung cancer screening, and (4) physicians' perspectives on the advantages and disadvantages of using CT scans as a screening test for lung cancer. The focus groups were assembled and audio-recorded by a commercial telephone conferencing service. The protocol for this study was reviewed and approved by the institutional review boards of CDC and RTI. All participants provided written informed consent and received $175 for their participation.


We compiled verbatim transcripts for each group and entered the transcripts into ATLAS.ti, (version 5.6.1)—a qualitative analysis program. Using the constant comparative method of qualitative analysis,[14] we created and defined inductive codes as concepts emerged from the data. We further examined the transcripts to identify novel themes and expand upon existing themes, refining the codes as appropriate until we reached a final set of codes. We then developed a detailed codebook following a standard structure.[15] Using the codebook, two researchers jointly coded all five transcripts, resolving any conflicts in order to achieve consensus. Inter-rater reliability was not calculated due to the joint coding approach employed. Once coding was complete, tables were developed to further analyze specific themes.[16]