Logo of hsresearchLink to Publisher's site
Health Serv Res. Aug 2005; 40(4): 953–956.
PMCID: PMC1361197

The Persistent Challenge of Avoidable Hospitalizations

Carolyn M Clancy, Agency for Healthcare and Research Quality

In the late 1980s John Billings developed the concept of counting potentially avoidable—or ambulatory care sensitive—hospitalizations as an indirect reflection of problems with access to care and effective primary care. (Billings and Teicholz 1990) The appeal of this approach is self-evident: the majority of states have good data on hospital discharges and limited data about ambulatory care. Avoiding a hospital admission represents a substantial “win” in restraining costs as well as enhancing patients' quality of life. Since that initial study, a body of literature examining avoidable hospitalizations has emerged, demonstrating a higher rate of such admissions among the poor and less well educated in many urban areas. As the U.S. struggles to increase the return on investment in public programs, particularly the Medicaid program, identifying mutable factors associated with avoidable hospitalizations is an issue of considerable urgency to policy makers and the public.

In this issue of the journal, two papers focus on factors impacting rates of avoidable hospitalizations. Leslie Roos et al. examined rates of ambulatory care sensitive hospitalizations for urban residents in Manitoba between 1998 and 2001. (Roos et al. 2005) Using both physician claims and hospital discharge abstracts, they found that residents from the lowest income neighborhoods had higher rates of hospitalizations than their counterparts in higher income areas. However, these same residents also had higher utilization of physician visits as well. Earlier studies by the Roos team have not identified specific factors related to compliance, individual characteristics (e.g., homelessness) or severity that can explain the findings presented here. The authors conclude that doing “more of the same” (e.g., increasing physician supply) is unlikely to have a significant impact on hospitalization rates.

In contrast, Laditka and colleagues used a national dataset to examine rates of ambulatory care sensitive hospitalizations in 642 urban counties and 306 rural counties in the U.S. (Laditka, Laditka, and Probst 2005) The team used multiple ecological variables to control for intercounty differences in race, ethnicity, air quality and health system use and characteristics. They find that physician supply is inversely correlated with rates of ambulatory care sensitive hospitalizations in urban areas but had no effect in rural areas. Of note, available data limit their capacity to examine important characteristics within individual counties or physician visits.

Assuming a correlation between physician supply and use of physician services, these apparently divergent findings with important policy implications warrant elaboration. Notwithstanding important differences in health care financing in the U.S. and Canada, is it plausible that physicians are so much more effective south of our shared border—or should we rightly conclude that the literature on ambulatory care sensitive hospitalizations represents an unfinished chapter in health services research, that is, a puzzle with missing pieces?

The premise underlying the concept of ambulatory care sensitive hospitalizations derives from a plausible expectation that timely access to effective primary care can result in decreased rates of admissions for selected conditions known to have high rates of variation. Of note, this hypothesis was tested in a randomized controlled trial. Weinberger et al. conducted a multicenter trial involving nine Veterans Affairs Medical Centers in which 1396 veterans hospitalized with one of three chronic conditions represented in most lists of potentially avoidable admissions (congestive heart failure, diabetes mellitus, and chronic obstructive pulmonary disease) were randomized to either usual care or an intensive primary care intervention. (Weinberger, Oddone, and Henderson 1996), This study, restricted to veterans who received all of their care within the VA system, found that veterans in the intensive primary care intervention group were rehospitalized more frequently than those in the usual care group. Veterans in the intervention group were more satisfied with their care, but had no differences in overall quality of life scores compared with those in the usual care group. The results prompted important questions about the content of primary care as well as unmet needs among a severely ill group of patients.

Health services researchers are all too familiar with the strengths and limitations of using data collected for one purpose (billing) to glean insights about the content of care provided. Neither study provided information on emergency department utilization, and other studies have shown inaccuracies related to identifying patients' insurance status, particularly those insured by Medicaid. (Chattopadhyay and Bindman 2005). Continued enhancements to administrative data could improve our capacity to identify both the patient and community characteristics associated with increased rates of ambulatory care sensitive hospitalizations and inform targeted efforts to improve access to effective primary care. Such enhancements in the U.S. might include the capacity to examine patterns of care at a level lower than a county.

An alternative hypothesis is that we have not yet identified which components of primary care, including community supportive services, are most effective in helping individuals with chronic and acute conditions frequently associated with hospitalizations manage their care, especially those in lower socioeconomic groups. Laditka notes that federal programs such as the National Health Service Corps and state loan repayment programs have resulted in increased recruitment and retention of primary care physicians to underserved areas. Many of these physicians provide care in organizations such as community health centers that include the capacity to mitigate multiple access barriers confronting poor populations, such as transportation, translation, and low health literacy. The data provide little information on the settings in which these physicians provide care, and the potential impact of additional services.

In fact, we have a great deal to learn about the components of effective primary care for all patients, irrespective of income, education or insurance status. For example, the work of Casalino et al. (2003) found that even organized medical groups with 20 or more physicians do not always have effective care management processes. In addition, multiple studies of quality care confirm a substantial gap between the best care and that which is routinely provided. For example, AHRQ's National Healthcare Quality Report found that only 32 percent of adults with diabetes have received five recommended services in the previous 2 years. (AHRQ 2005)

Continued efforts to improve the accuracy and timeliness of administrative data can help pinpoint areas where improvements in access to effective primary care are most important. Such enhancements should not obscure our limited understanding of which interventions are most effective in preventing avoidable hospitalizations, particularly for vulnerable populations. The time to move beyond generating hypotheses and questions to evaluation of organizational and clinical strategies that promote access to high quality primary care is long overdue.


  • AHRQ “National Healthcare Quality Report.” 2005. Available at http://www.qualitytools.ahrq.gov(accessed June 1, 2005).
  • Billings J, Teicholz N. Uninsured patients in District of Columbia hospitals. Health Affairs (Millwood) 1990;9(4):158–65. [PubMed]
  • Casalino L, Gillies RR, Shortell SM, Schmittdiel JA, Bodenheimer T, Robinson JC, Rundall T, Oswald N, Schauffler H, Wang MC. “External Incentives, Information Technology, and Organized Processes to Improve Health Care Quality for Patients with Chronic Diseases.” Journal of the American Medical Association. 2003;289(4):434–41. [PubMed]
  • Chattopadhyay A, Bindman AB. “Accuracy of Medicaid Payer Coding in Hospital Patient Discharge Data: Implications for Medicaid Policy Evaluation.” Medical Care. 2005;43:586–91. [PubMed]
  • Laditka JN, Laditka SB, Probst J. “More May Be Better: Evidence of a Negative Relationship between Physician Supply and Hospitalization for Ambulatory Care Sensitive Concerns.” Health Services Research. 2005;40(4):1148–1166. [PMC free article] [PubMed]
  • Roos LL, Walld R, Uhanova J, Bond R. “Physician Visits, Hospitalizations, and Socioeconomic Status: Ambulatory Care Sensitive Conditions in a Canadian Setting.” Health Services Research. 2005;40(5):1167–1185. [PMC free article] [PubMed]
  • Weinberger M, Oddone EZ, Henderson WG. “Does Increased Access To Primary Care Reduce Hospital Readmissions? Veterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission.” New England Journal of Medicine. 1996;334(22):1441–7. [PubMed]

Articles from Health Services Research are provided here courtesy of Health Research & Educational Trust
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...