Rates of Preventable Hospitalization for Ambulatory Care Sensitive Conditions (ACSCs) among Patients with Lymphoid Malignancies (LM)

BACKGROUND: The influence of comorbid medical illness on treatment outcome and survival from LM has been well-characterized. Recent reports suggest that optimal management of these comorbidities may also be important. We sought to indirectly determine the effectiveness of outpatient treatment for ACSCs, conditions where good outpatient care can potentially prevent the need for hospitalization, by calculating population-based estimates of hospital admission rates among patients with LM.

Methods: Data were obtained from the 2005 Nationwide Inpatient Sample. Using ICD-9CM codes, we identified all adult (age ≥20) admissions to U.S. community hospitals for LM (Hodgkin’s disease, non-Hodgkin’s lymphoma, and multiple myeloma). A comparator group without known diagnosis of cancer was created by excluding records containing any diagnosis code for malignant neoplasm or diagnosis/procedure code for cancer treatment. ACSC admissions, including those for short- and long-term complications of diabetes mellitus (DM), uncontrolled DM, asthma, hypertension (HTN), congestive heart failure (CHF), angina, and hypovolemia, were ascertained using algorithms developed and validated for the U.S. Agency for Health Care Research Quality Prevention Quality Indicators. The 2005 5-year prevalence for LM was obtained from SEER and used as the denominator for rate calculations in that group. A denominator for the no cancer group was created using U.S. Census estimates for the 2005 adult population less the SEER 5-year prevalence for all sites. Mean hospital charges were extracted for each admission and transformed into costs using Medicare cost-to-charge ratios. Length of stay, total costs, and in-hospital mortality were compared across groups for each ACSC. All means and proportions were sample weighted.

Results: In 2005 there were an estimated 510,300 total LM admissions and 26,700,000 total admissions in the no cancer comparator group. Estimated hospitalization rates for each ASCS and odds ratios for the between group comparisons are detailed below.

In-hospital mortality did not significantly differ between groups for any ACSC. Mean length of stay and hospital costs were likewise similar with the exception of costs for CHF ($8,957[95%CI 8,260–9,654] v. $7,176 [6,185–8,168]) and length of stay (5.6d [95%CI 4.8–6.3] v. 4.0d [3.9–4.1]) and costs ($8,702 [6,832–10,572] v. $5690[5,373–6,007]) for asthma admissions.

Conclusions: Hospitalization of LM patients for ASCSs is common and occurs with odds generally >2 times higher than among patients without a cancer diagnosis. Future studies should be conducted to determine factors influencing these findings (e.g. rates of comorbidity, influence of cancer treatment, utilization of primary care services) and to develop potential strategies for preventing hospital admissions.