| OCR Text |
Show been proposed as indicators of access to primary care (Bindman et al, 1995; Hrdy et al., 1993; Millman, 1993). Access is "the timely use of personal health services to achieve the best possible health outcomes" (Millman, 1993). Small areas with high ACSC hospitalization rates relative to surrounding areas thus become the focus of interventions to improve community health status and reduce costs through programs to address the causes of these hospitalizations (Griswold et al., 1992). Rural areas, defined by the U.S. Census Bureau as having 6-100 people/square mile, tend to have fewer resident health care providers than urban areas This situation is accentuated in frontier areas, with less than 6 people /square mile. Over 90% of frontier counties are located in the western U.S. Silver et al. (1998) identified a significantly greater rate of hospitalization in rural/frontier vs. urban Utah for elderly patients with acute and chronic ACSCs. Access itself is a composite of other variables, and ACSC hospitalization rates are potentially influenced by additional factors including local disease prevalence, severity of the case, patient attitudes, care-seeking behaviors, and physician practice styles (Billings et al., 1993; Gill, 1996; Grumbach et al., 1995; Hrdy et al., 1993; Komaromy et al., 1996; Parchman, 1995; Shea et al., 1992; Weissman et al., 1992), as well as random fluctuations, especially when the population base is very small (Parchman, 1995). Several studies have attempted to validate the use of ACSC hospitalization rates as a proxy for access, and have found correlations with self-reported access (Bindman et al., 1995). Other related factors include socio-economic status, both individually and as a contextual variable, race/ethnicity (Billings et al.; Bindman et al., 1995), regular source of care (Bindman et al., 1995), delay in care-seeking (Hrdy et al., 1993), supply of primary care providers (Bindman et al., 1995), and supply of family physicians (Parchman, 1995). Bindman (1995), Grumbach (1995), Komaromy (1996) and their colleagues studied non-medical influences on the physician's decision to admit a patient to the hospital for an ACSC in urban California. They used a mailed survey including hypothetical vignettes to demonstrate variation among urban California physicians' self-ratings of propensity to admit patients of comparable severity if patients were in situations with no doctor for follow-up, substance abuse, known to be non-compliant, inability to pay, homeless or living alone, or on Medicaid. They also asked patients about their perceptions of access in a second mailed survey. However, they did not find a significant correlation of practice style with ACSC hospitalizations once patient perceptions of access and other factors were controlled. The present study sought to develop possible explanations for the variation in Utah ACS hospitalizations of elderly, which was seen by Silver et al. (1998), by asking Utah physicians about their perceptions of the impact of specific non-medical factors on the physician's decision to admit a patient to the hospital with an ACSC, and whether non-medical factors might differentially influence this decision for urban vs. rural physicians. Methods Rosters of physicians in family practice, general practice, general internal medicine, and emergency medicine in each Utah county were developed based on information from the Utah Medical Association, the State Health Department, and telephone books. 78 urban and 111 rural/frontier Utah physicians were selected for a mailed survey using computerized randomization. The survey was modeled after one by Komaromy, Grumbach, and Bindman (1996), but major items were changed to reflect issues relevant to Utah physicians. Items were reviewed for face validity by a panel of physician experts. The survey was pretested by convenience panels of urban and rural physicians. Figure 1 displays the posed scenarios. They included one chronic condition of the elderly, congestive heart failure (CHF), as well as two acute conditions, pneumonia and cellulitis, which show considerable response to ambulatory care. First, subjects self-rated the decision to hospitalize patients in the scenarios based on medical reasons only (1=very unlikely, 2=unlikely, 3=maybe, 4=likely, 5=very likely). Then subjects rated the degree to which that decision would be influenced by each of the additional issues: 1. If you knew that this patient had excellent family support at home (spouse and extended family). 2. If you knew that this patient lived alone and had no close family or friends for support. 3. If you knew that this patient had excellent health insurance, in addition to his/her Medicare, that would cover his/her therapy (whether inpatient or outpatient). 4. If you knew that this patient had no health insurance beyond his/her Medicare, and that his/her income was at or below the poverty level. 5. If you knew that this patient lived very close to both the hospital and your office (less than 15 minutes away). 6. If you knew that this patient lived quite far from both the hospital and your office (at least an hour's drive). 7. If you knew that the patient's family was strongly in 20 |
