Published Online: May 10, 2012
Gery P. Guy Jr, PhD, MPH; and Lisa C. Richardson, MD, MPH
This article was published as part of a special joint issue and also appears in the Journal of Oncology Practice.
Objectives: To examine the characteristics of patients with cancer and their visits to outpatient, office-based physicians; to analyze any differences between visits to oncologists and visits to other physicians; and to examine the effect of patient, practice, visit, and geographic characteristics on the length of time patients with cancer spend with physicians during office-based visits.
Methods: We examined a total of 2470 patient office visits to nonfederally employed physicians from the 2006 and 2007 National Ambulatory Medical Care Survey. We performed descriptive analyses to examine the characteristics of patients with cancer by physician specialty. We conducted multivariate analyses using a generalized linear model to examine the relationship between visit duration and patient, practice, visit, and geographic characteristics.
Results: Forty-two percent of patients with cancer visited an oncologist. Females, females diagnosed with breast cancer, and individuals with advanced-stage cancer were more likely to visit an oncologist. Patients who visited oncologists were more likely to receive an anticancer drug, radiation therapy, and an increased number of diagnostic/screening services than those visiting other physicians. The mean duration of patient visits was 22.9 minutes. Higher percentages of performance-based compensation and capitation rates were associated with visits 4.4 minutes and 5.7 minutes shorter, respectively.
Conclusions: Higher use of performance-based payment mechanisms and capitated arrangements is associated with a decrease in the amount of time physicians spend with their patients with cancer. It is unclear whether shorter visit times impact the quality of medical care provided or whether physicians in these settings have become more proficient in caring for their patients.
(Am J Manag Care. 2012;18(5 Spec No. 2):SP49-SP56)
In 2007, more than 11 million Americans were living with cancer.1 It is estimated that up to $207 billion will be spent on cancer care annually by 2020.2 Given the high prevalence of cancer and the substantial resources devoted to its treatment, it is important to understand the characteristics of medical care use among patients with cancer. A recent study showed that, from 2002 to 2003, the overwhelming majority of patients with cancer (87%) visited physician offices; the remaining visited hospital clinics.3
Among the important characteristics of outpatient visits are the type of physicians seen and the length of time patients spend with their physicians. Many factors have been shown to affect outpatient visit duration, such as patient characteristics, practice characteristics, insurance coverage, and visit content.4,5 The length of time a physician spends with a patient is strongly associated with physician productivity, which directly affects the number of patients seen per day and potentially affects practice revenue.4
There is a clear trade-off between physician productivity and the quality of care provided. Although reducing visit length can improve physician productivity, it also has the potential to reduce the quality of care provided. For example, shorter outpatient visits have been shown to be associated with decreased patient satisfaction and trust,6-9 less attention to patients’ psychosocial problems,10 and a decrease in the provision of certain preventive health services.11,12 In turn, low patient satisfaction has been associated with poor treatment adherence13 and worse chronic disease outcomes.14 These factors are especially important among those, such as patients with cancer, with health conditions requiring long-term treatment and follow-up.
In this study, we examine the characteristics of patients with cancer seen by outpatient office-based physicians. Specifically, we examine the characteristics of patients seen by oncologists compared with those seen by other physicians. Additionally, we analyze the effect of patient, practice, visit, and geographic characteristics on the length of time patients with cancer spend with physicians during outpatient visits.
We analyzed data from the 2006 and 2007 National Ambulatory Medical Care Survey (NAMCS). The NAMCS is a national probability sample survey of visits to nonfederal office-based physicians conducted by the Centers for Disease Control and Prevention’s National Center for Health Statistics. The survey uses a multistage probability design involving samples of geographic primary sampling units, physician practices within primary sampling units, and patient visits within physician practices. Physicians were identified from the master files of the American Medical Association and the American Osteopathic Association. In 2006 and 2007, the NAMCS oversampled oncologists. Sampled physicians were asked to complete patient record forms for a systematic random sample of approximately 30 office visits during a randomly assigned 1-week period.15,16
In 2006, 64% of eligible physicians agreed to participate and 65% participated in 2007, resulting in 3023 physicians reporting on 62,170 visits.15,16 In 2006 and 2007, response rates among oncologists were 45% and 56%, respectively, resulting in 120 oncologists reporting on 2734 visits.15,16 To study the characteristics of outpatient visits for cancer, we limited the analysis to all office visits with the primary diagnosis of cancer, coded according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). Cancer type was categorized as follows: larynx/lung (ICD-9-CM, 161-162), female breast (174), prostate (185), colon/
rectum (153,154), leukemia/lymphoma (200-208), melanoma (172), and other malignancies. Primary diagnoses of nonmelanoma skin cancer were excluded (ICD-9-CM, 173).17 Whereas there were 2559 eligible visits, 89 visits (3.5%) were excluded from our analyses because the patients did not have face-to-face contact with the physicians.
For each visit included in the NAMCS, the physician, a staff member, or a US Census Bureau field representative recorded information on patient characteristics, practice characteristics, visit content, and geographic characteristics. The NAMCS contains data on patient demographics, health insurance, physician specialty, reason for the visit, cancer site and stage, medications ordered or provided, diagnostic/screening services ordered or provided, physician compensation mechanisms, and geographic characteristics. The key variable of interest in our analysis is the time spent with the physician, a continuous variable measuring the face-to-face interaction time in minutes between the physician and patient. Time spent waiting for the physician or seeing other healthcare providers is not included in the measure of visit duration.
We conceptualized that the duration of a visit is a function of patient characteristics, practice characteristics, visit characteristics, and geographic characteristics. Patient characteristics include age, sex, race/ethnicity, type of health insurance, whether the patient was new to the physician, whether the patient was referred, the presence of other chronic conditions, and cancer type and stage. Practice characteristics include physician specialty and practice ownership status. Several measures of physician compensation are also included—whether productivity, patient satisfaction, quality of care, or practice profiling is taken into account when determining patient-care compensation. Additionally, the percentage of patient-care revenue on the basis of bonuses, returned withholds, or other performance-based payments and the percentage of revenue from capitation were examined. Visit characteristics include the reason for the visit, number of diagnostic/screening services ordered or provided,
number of health education services ordered or provided, whether radiation therapy was ordered or provided, whether an anticancer medication (antineoplastics) was administered or ordered, and visit disposition. Geographic characteristics include whether the physician practices in a standard metropolitan statistical area, region of the practice location (East, Northeast, Midwest, or West), and socioeconomic indicators (percent poverty and percent of adults with a bachelor’s degree or higher) at the patient zip code level.
Descriptive analyses were performed to examine the distribution of patients by cancer type and physician specialty. Additionally, descriptive analyses were performed to examine patient characteristics, practice characteristics, visit characteristics, and geographic characteristics stratified by the oncology specialty. We conducted a bivariate analysis to examine the differences between patients seen by oncologists and those seen by other physicians. Categorical variables were analyzed using Pearson’s χ2 test, whereas continuous variables were tested by using analysis of variance.
To examine the factors influencing duration of visits by patients with cancer, we first examined the mean and standard errors (SEs) of visit duration across various characteristics. The relationship between visit duration and these characteristics was then modeled, controlling for patient, practice, visit, and geographic characteristics. Multivariate analyses were performed by using a generalized linear model with a gamma distribution and a log link to account for the skewness of the visit duration variable. The marginal effects of each of the explanatory variables on the dependent variable are reported and can be interpreted as the change in minutes associated with each explanatory variable, independent of the other variables.
To examine the sensitivity of our findings to the specification of our dependent variable, we also analyzed our data using a multivariate logistic regression model with a dependent variable indicating whether visit duration was longer than 30 minutes. Overall, results of this analysis closely resemble the results from the generalized linear model, which suggests that the findings are not sensitive to the specification of the dependent variable. All analyses were conducted using survey data commands in Stata version 11.2. Patient characteristics were examined using patient weights, whereas visit length was examined using visit weights.18
In 2006 and 2007, NAMCS sampled data on 2470 office visits for cancer, representing a total of 10.7 million patients. Table 1 presents the distribution of patients by cancer type and physician specialty. Overall, 41.9% of patients were seen by oncologists, 15.0% by urologists, 7.2% by hematologists, 3.8% by dermatologists, 10.8% by general surgeons, and 9.8% by primary care providers. The majority of patients with female breast (60.1%), colorectal (54.5%), and lymphoma/ leukemia (55.9%) cancer were seen by oncologists, whereas patients with prostate cancer were primarily seen by urologists (69.7%) and patients with melanoma were primarily seen by dermatologists (64.7%).
PDF is available on the last page.