Factors associated with physician perceptions of and willingness to recommend inhaled insulin

To examine predictors of physician perceptions of an inhaled insulin, willingness to prescribe that insulin, and estimates of patient initiation of therapy with that insulin. The study was an Internet survey of a US national sample (n = 602) of physicians who treat adults with diabetes. Respondents were given a brief description of potential clinical benefits and administration procedures for the study inhaled insulin (SII). Measures included clinical beliefs, benefit ratings and overall evaluation of SII relative to other mealtime insulins, willingness to recommend SII, and estimates of patient initiation of therapy with that insulin. Multivariate regression assessed significant independent associations controlling for respondent and patient case-mix characteristics. Physicians who self-identified as medical innovators or who reported high levels of involvement with patients tended to rate the SII higher, while respondents who self-identified as diabetes experts or who avoided using insulin tended to rate the SII lower. Medical innovators and those who rated the SII high on efficacy in avoiding discomfort and inconvenience were more likely to say they would recommend the SII to their patients and that their patients would use it. Family physicians were most likely and endocrinologists least likely to say they would recommend the SII for a variety of patient profiles. Physicians see a variety of important benefits for the SII, and would recommend inhaled insulin to patients with different treatment regimens and treatment needs, especially those patients who are hesitant to initiate insulin therapy or concerned about taking more insulin injections. These findings should be considered in light of study limitations, including the fact that responses were based on expected benefits, and not on benefits actually experienced by physicians in the study, the fact that no information was provided about the cost of the SII, though this could have an important influence on prescription decisions, and the fact that the study sample was a self-selected group, rather than a representative sample of all physicians treating patients with diabetes.