Table 1 Multilevel Model Variables: Description, Rationale, Source, and Sample Statistics
Variable (units of measure) | Rationale for Inclusion | Source, date | BC Models | CRC Models | ||
|---|---|---|---|---|---|---|
Outcome: whether cancer patient was diagnosed at a late stage (regional or distant =1, else = 0) | Late stage diagnosis is indicative of lack of knowledge regarding personal cancer risk, or the importance or availability of screening; lack of timely or proximate access to services, lack of funds to pay for, and cultural or other barriers related to utilization of timely cancer screening. | SEER and NPCR cancer registry data made available through NCHS Research Data Centers, covering 2004–2009: http://www.cdc.gov/rdc | mean | sdev | mean | sdev |
0.308 | 0.461 | 0.543 | 0.498 | |||
Person-level predictors | ||||||
female (binary) | Only females are included in BC study. Although males do have BC incidence, the numbers are few. Both male and female are included in the CRC study, with male designated as the reference group. | 1.000 | 0.000 | 0.487 | 0.500 | |
black (binary) | The national statistics cite blacks as a disadvantaged group, with worse outcomes relative to whites, the reference group. | 0.101 | 0.301 | 0.112 | 0.315 | |
race all other (binary) | All other races and ethnicities were combined to make the model more parsimonious, relative to whites, the reference group. Includes 8 % Hispanic, 3 % Asian, 0.5 % Native American, 0.8 % other. | BC sample n = 981,457 CRC sample n = 558,568 | 0.126 | 0.332 | 0.124 | 0.329 |
age < 65 (binary) | Two age groups allow us to distinguish effects for the well-insured Medicare population from the less well-insured younger population with cancer, who may also be more genetically susceptible and more likely to be screened and diagnosed at late stage. | 0.624 | 0.505 | 0.424 | 0.494 | |
County-level predictors | ||||||
isolation black (index 0–1) | Isolation indices have been examined in a broad literature as contextual predictors of health behaviors and outcomes. At smaller geographic scales they are thought to represent social support, and at broader scales political clout. A higher index value represents a lower chance that minorities reside among whites, with a value of 1 indicating a perfectly segregated society (2000). | RTI Spatial Database (https://rtispatialdata.rti.org) which includes variables derived by authors under NIH funding and made publicly/freely available. Sources include Decennial Census 2000; CMS Geographic Area Service Files, 2005;100 % FFS Medicare claims extracts 2006; Census SAHIE 2005 | 0.257 | 0.214 | 0.259 | 0.217 |
isolation Asian (index 0–1) | 0.072 | 0.086 | 0.068 | 0.085 | ||
isolation Hispanic (index 0–1) | 0.216 | 0.203 | 0.209 | 0.205 | ||
managed care penetration (%) | Managed care has transformed the way medicine is practiced in highly-penetrated markets, with preventive care services more prevalent/utilized more intensively (2005). | 15.9 | 14.7 | 15.3 | 14.7 | |
Distance (miles) | Calculated as the average distance (miles) over all ZIP codes with centroid in the county to closest provider ZIP code. Greater distance to provider of BC (mammogram) or CRC (endoscopy) screening suggests impeded access to preventive care services. Based on 100 % FFS Medicare utilization of mammography or endoscopy services (2006). | 6.02 | 6.10 | 5.15 | 4.80 | |
Screening rate (%) | Percent of the 100 % FFS Medicare population residing in the county and alive all year that utilized cancer screening (mammography, endoscopy) (2006). | 23.60 | 3.18 | 11.05 | 1.43 | |
Percent uninsured (%) | % of the under-age-65 population who did not have health insurance (2005). | 17.73 | 5.45 | 17.75 | 5.49 | |
State-level Policy Variables | ||||||
Direct Access to Specialist (1 = yes, 0 = no) in 2004 | Access to gastroenterologists, gynecologists or oncologists without need of referral from a primary care physician may result in better matching of patient/provider and more timely care. Hypothesized to increase access for less well insured individuals or those in more stringent managed care plans. Younger people tend to be enrolled in these plans, which are less costly but restrict access and choice. Source: NCSL, 2010. | 0.956 | 0.206 | 0.951 | 0.216 | |
Ban on Financial Incentives (1 = yes, 0 = no) in 2004 | Not allowing insurers to reward physicians financially for substitution of lower cost care could result in use of more expensive cancer screening tests (e.g., colonoscopy vs sigmoidoscopy or FOBT), more accurate surveillance and better quality care. For BC screening, this law could impact prescribing the more expensive MRI breast exam recommended for denser breast tissue versus mammogram. Source: NCSL, 2010. | 0.628 | 0.483 | 0.627 | 0.484 | |
Greater Practice Latitude for Nurse Practitioners (1 = yes, 0 = no) in 2004 | Allowing nurse practitioners the latitude to practice medicine in independent clinics, without physician supervision can improve access to primary care in underserved areas. Hypothesized to increase access to primary care providers, increasing the chance that a person will be encouraged to utilize cancer screening. Source: NCSL, 2013. | 0.342 | 0.474 | 0.336 | 0.472 | |