Abstract
In response to our extended critiques (Hickie et al., 2011; Rosenberg and Hickie, 2010) of the woefully inadequate evaluation of the Better Access initiative, Pirkis et al. (2011b) perpetuate key myths about our earlier very large SPHERE studies [386 general practitioners (GPs) and 46,515 patients]. First, they state we used the method of ‘. . . recruitment of patients by providers . . .’ (Pirkis et al., 2011b). While patients were recruited in general practice settings, they were not selected by the GPs. Instead they were recruited by administrative staff in waiting rooms. Patients and GPs then completed key information independently with regards to their mental health status and the treatments provided. We included in our analyses (Hickie et al., 2001b) only GPs who provided data on more than 50 patients (mean number of patients screened per GP = 129, range 50–435). Our method stands in marked contrast to the very limited, provider-driven selection and reporting of data used in the evaluation of Better Access (5–10 patients per provider, 129 providers, 883 patients assessed) (Pirkis et al., 2011a). Second, Pirkis et al. (2011b) refer to the representativeness of each sample with regards to providers. As we made clear, we used the national framework of available providers at the time (N = 22,000 GPs). We also did not use data from those GPs who responded, but provided data on less than 50 patients screened. To compare like with like, our actual response rate would be 418/22,000 (1.9%) and Pirkis et al. (2011c) would have 129/49,000 [0.3%; or, taking into account available practitioners, 39/24,000 (0.2%) GPs and 90/25,000 (0.4%) allied health professionals]. That is, while all such studies only utilize a very small proportion of the available workforce, the key issues with regards to detailed analysis of who receives care, and the type of care provided, depends on the actual number of patients assessed (i.e. SPHERE N = 46,515 vs. Better Access N = 883).
We agree that substantive effectiveness and related high-quality health services research should underpin major health system developments. Sadly, the very limited scope of the evaluation of Better Access (Pirkis et al., 2011c) prevented a genuine assessment of the key issues that still surround the initiative (Hickie et al., 2011), namely: who is receiving mental health services in primary care (i.e. age, gender, socioeconomic status, geography, non-English speaking background, severity of illness); who is providing those services (GPs alone as compared to collaborative care with other trained mental health professionals); the experiences of care of those receiving and not receiving such services; and, are the additional services being provided in the most cost-effective manner? Contrary to the assertion of Pirkis et al. (2011b), we are applying exactly those standards to the evaluation of the headspace programs with which we are associated (Hamilton et al., 2011; Scott et al., 2009, 2012). The key issues do not, and never did, relate to the effectiveness of psychological interventions in primary care settings.
We continue to assert that for an initiative that has provided more than 11.1 million mental health services to over 2 million people since its inception in 2007 (The Senate – Community Affairs Reference Committee, 2011), analysis of a cohort of less than 1000 highly-selected cases was grossly inadequate. It has left the field (and the Government) without the necessary information on which to base future key policy decisions. This is particularly true when one reflects on the size and scope of what can actually be done in services evaluation in primary care settings in Australia (Davenport et al., 2001; Hickie et al., 2001a, 2001b). We note that the recent independent Senate Inquiry has just arrived at the same conclusion (The Senate – Community Affairs Reference Committee, 2011).
See Original Article by Pirkis et al., 2011, 45(9): 726–739; See also Editorial by Pirkis et al., 45(11): 911–914