Documentation of the 5 As for Smoking Cessation by PCPs Across Distinct Health Systems
Published Online: March 14, 2014
Rebecca J. Williams, DrPh; Andrew L. Masica, MD; Mary Ann McBurnie, PhD; Leif I. Solberg, MD; Steffani R. Bailey, PhD; Brian Hazlehurst, PhD; Stephen E. Kurtz, PhD; Andrew E. Williams, PhD; Jon E. Puro, MHA-PA; and Victor J. Stevens, PhD
Tobacco use remains the leading cause of preventable death in the United States, and yet nearly 46 million people currently smoke cigarettes.1-3 Primary care physicians can greatly facilitate smoking cessation among their patients, and this effort can be a cost-saving clinical preventive service.4-6 Well-accepted, evidencebased guidelines for delivering tobacco-cessation treatments in the primary care setting have been shown to be effective and are recommended as standard care.7-13 Physicians are encouraged to help patients quit by implementing the 5 As of smoking cessation, which include (1) asking all patients about their smoking status; (2) providing personalized advice to quit; (3) assessing smokers’ readiness to quit; (4) assisting motivated patients in their quit attempts, and (5) arranging follow-up contacts for those receiving assistance to quit.
Despite these recommendations and the fact that almost 70% of smokers are interested in quitting,14 the overall proportion of patients who receive counseling remains low15 and there is substantial variation in the use of the 5 As framework across clinical practices and demographic groups.7,14,16,17 Provision of the first 2 As (ask and advise) has been steadily increasing,8,18 but delivery of assess, assist, and arrange has been inconsistent.7,8,10,12,15,19 The rate of cessation help is often lower for Hispanic patients than for non-Hispanic whites; and younger smokers are less likely to receive cessation assistance compared with older smokers.20
Little is known about adherence to the 5 As since the 2008 publication of updated recommendations on treating tobacco use and dependence. 7,8,10,12,19 Prior studies assessing provision of the 5 As have generally relied on patient questionnaires, or data collection was limited to specific coded fields within patient medical records.8-10,12,19 Additionally, prior work in assessing the 5 As has been done within a single health system, limiting feasibility of extending findings to other health systems.
This paper extends previous work by leveraging comprehensive medical record data, including coded data fields (eg, diagnosis codes), progress notes, and patient educational materials about smoking, which were handed out to the patient during a visit to assess implementation of the 5 As. A random sample of 200 smokers from each of 6 different health systems across the United States was used. The data reported here were generated by a manual chart review process involving the Comparative Effectiveness Research Hub (CER Hub), a network of asthma and smoking cessation researchers from each of the 6 health systems used for this study (www.cerhub.org).21 Smoking cessation activities based on the 5 As were implemented within the 6 health systems that collectively make up the CER Hub.
The study aims are 3-fold: (1) identify the proportion of known smokers that receive smoking cessation services in the course of routine clinical practice; (2) describe demographic and comorbidity characteristics of smokers receiving smoking cessation services in these systems; (3) evaluate differences in performance of the 5 As across health systems, and also across gender and age categories. Health systems could benefit from an analysis of the 5 As, using patient records to assess their own performance and to guide future smoking cessation initiatives.
Data were collected from 6 health systems, representing unique demographic and geographic populations. The 6 sites constitute a convenience sample of medium-sized healthcare organizations. While this group cannot be seen as a representative sample for the United States, this group does have some geographic diversity, and represents a fairly diverse patient population in terms of race, ethnicity, and socioeconomic status. Each health system also has a distinctive technology infrastructure and electronic medical record (EMR) system. The Institutional Review Board at each health system approved the study protocol. Descriptions of participating health systems are provided in Table 1.
The medical records of 200 current smokers in each of the 6 selected health systems were randomly selected for this study (N = 1200). Inclusion criteria included patients 12 years and older who were identified as (1) a current smoker at any point during the years 2006 to 2010, regardless of “quitting” events and (2) having at least 1 primary care visit in each of the 5 study years. We classified patients as “current smokers” if (1) they received any of the following International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9- CM) diagnosis codes related to tobacco-use disorders: 305.1, 649.01, 649.02, 649.03, 649.04, or 989.84, at any primary care visit or (2) had an update to their social history indicating “current smoker” during the study period. Smoking-related comorbidities during the observation period were also documented for each participant, and ethnicity and race information was extracted from each patient’s medical record.
Coding of the 5 As
A single primary-care encounter from each of the 1200 medical records, as described above, was hand coded for documentation of the 5 As. An “encounter” included all data linked to the encounter (eg, vital signs, reasons for visit, medications and procedures ordered, diagnoses) as well as progress notes and patient education materials printed out for the patient. Encounter records were formatted as XML computer language files according to a schema conforming to Health Level Seven International’s (HL7’s) Clinical Document Architecture (CDA) standard.22,23 These records were de-identified and uploaded onto a secure Web server, where they were accessed by project staff using a Web-based application called Manual Coder that is part of the CER Hub. Users of Manual Coder sequence through encounter records in a data set and have the ability to highlight text elements representative of the study measures (in this instance, components of the 5 As) and then select the code appropriate to the highlighted text from a list. The applied codes, as well as the highlighted content from the encounter record, can be extracted from the database for analysis of either site-specific or aggregate data.
PDF is available on the last page.