Objective: To assess early patient dropout rates during infertility treatment as a potential measure of wasted resources. Methods: The study involved multifaceted population cohorts, including a prospectively observed captive health maintenance organization (HMO) population and retrospectively selected preferred provider organization (PPO) patients. One hundred twenty-eight HMO couples were followed prospectively for 6 months. The insurance carrier retroactively selected 96 couples from their PPO population who were believed to be infertility patients. They were matched by date, age, and time of hysterosalpingography to infertility patients in the carrier's HMO population. Patients were considered treatment dropouts if they either requested their provider to abandon further work-up or treatment, or if they failed to return for an appointment for 3 months. Results: Forty-six of 128 (36%) HMO patients followed prospectively discontinued care within 180 days, with only eight (6.3%) providing defined reasons. Preferred provider organization patients uniformly demonstrated significantly higher dropout rates than HMO patients, a finding already apparent at 60 days (P <.002; odds ratio [OR] 3.67, 95% confidence interval [CI] 1.47–9.97) and 120 days of treatment (P =.002; OR 2.87, 95% CI 1.39–6.06). Among PPO patients, dropout rates were especially pronounced if infertility care was provided by generalists. At billing levels of at least $2000, HMO patients also demonstrated less dropout than PPO patients (P <.001; OR 6.14, 95% CI 2.72–14.79), with generalists again demonstrating a significantly larger patient loss than infertility specialists (P <.001; OR 0.18, 95% CI 0.66–0.49). Conclusion: Infertility patients demonstrate a surprisingly large early dropout rate, which is significantly larger if patients receive infertility care from generalists rather than specialists. Newly presenting infertility patients should be carefully evaluated, especially in indemnity situations, before expensive diagnostic and therapeutic interventions are ordered.