Provider Characteristics associated with HAART Initiation among an ARV Naïve Population

Objective: Potent antiretroviral therapy (HAART) is associated with increased survival among HIV-infected persons. Despite this benefit, little is known regarding whether certain provider level characteristics influence the likelihood of HAART receipt. This study had two objectives: 1) determine whether provider characteristics are associated with HAART receipt and 2) identify clinical and treatment regimen characteristics associated with HAART initiation, adjusting for provider characteristics.

Study Design: We retrospectively studied 2,069 anti-retroviral (ARV) naïve patients receiving care within Kaiser Permanente HMO in Northern California and Oregon. Eligible patients had a HIV RNA >75 copies/mL between 7/1/1999 and 6/30/2006 and were followed until receipt of HAART, disenrollment, death or end of follow-up. Provider characteristics included HIV panel size and years of HIV clinical experience. Clinical measures included CD4 count, HIV RNA level, and AIDS criteria (1993 CDC). Treatment regimen characteristics included total pill count and a measure that assessed when fixed-dose combination (FDC) treatment regimens were implemented into clinical practice (i.e. 1/1/2005 and beyond). Time-dependent Cox regression models were used to examine the independent effect of provider, clinical and treatment regimen characteristics associated with receipt of HAART, adjusting for age, gender, race/ethnicity and injection drug use status.

Results: The population was 88% male (median age = 39) with care provided by a total of 337 providers (median years of experience = 12). Patients receiving care from providers with 16 or more years of HIV experience were significantly more likely to receive HAART (HR=1.34, 95% CI = 1.04 – 1.72), compared to those with 0-5 years of experience; after adjusting for demographics, clinical and treatment regimen characteristics. The following characteristics were also significantly associated (p<0.05) with initiation of HAART: CD4 <=200, AIDS criteria, HIV RNA >= 100,000/mL and higher pill burden.

Conclusions: Years of HIV experience among practitioners significantly impacts the likelihood of HAART receipt. In addition, clinical characteristics and total pill burden appear to be important predictors of HAART receipt. Future research should include more comprehensive assessment of provider characteristics and examine receipt of HAART beyond HMO settings.