LAUSR

The article by Daskivich et al1 in this issue of JAMA Internal Medicine evaluates a large-scale telemedicine diabetic retinopathy (DR) screening program in the Los Angeles County Department of Health Services, one of the largest safety net health care systems in the United States. This widespread screening program used existing primary care workflows to train medical assistants and licensed vocational nurses to be the certified fundus photographers and then sent the digital images to staff optometrists to grade levels of DR and determine timing of follow-up appointments for appropriate specialty care. Improving screening of DR is a key public health priority, given that diabetes is the leading cause of preventable blindness in the United States. Safety net health systems have traditionally faced challenges conducting recommended annual screening for DR because of the high prevalence of patients with diabetes combined with the lack of access to optometrists and ophthalmologists. The authors review both the overall results of implementation at the system level, as well as the results of a smaller patient-level analysis of clinical changes among randomly selected individuals within the program, finding that the median time to screening for DR decreased substantially (from 158 to 17 days) and that overall rates of screening for DR increased from 39.8% to 55.4%, with a total of 21 122 patients screened across the system. On their own, these improvements in wait times and rates of screening for DR are laudable, given that specialty care has been historically underresourced within large safety net health care systems such as the Los Angeles County Department of Health Services.2 Herein, we highlight how Daskivich et al1 used several key implementation strategies to achieve these results. First, standardizing workflow for making referrals, and putting this standard work in the hands of nonphysician health care team members such as medical assistants, has been shown to improve the delivery of recommended care.3 Similarly, specialty care professionals have been shown to have varying levels of flexibility with regard to accepting referrals into their busy practices,4 and this program has standardized the work for optometrists to make triaging decisions about which patients should be seen, as well as how quickly they should be seen.1 Finally, because the lack of clear communication between primary and specialty care creates inefficient use patterns, the program used an existing electronic referral platform (similar to effective electronic referral systems used in other safety net health care systems)5 to communicate seamlessly between clinics about results of screening for DR and scheduling future appointments.1 Combining several evidence-based strategies for health system innovations has produced improved provision of care without large increases in cost. Although these implementation solutions seem straightforward and clear, they actually represent cultural shifts in work responsibilities, as well as expectations on the part of both primary care and specialty professionals and staff. This finding suggests that much of the innovation in this telemedicine DR screening program is not limited to the new fundus camera technology but can be found in the use of such technology in the context of several new team-based clinical workflows to create more efficient outcomes, which supports the findings of other previous studies on high-functioning health care systems.6 Daskivich et al1 state in several places that these workflows are multifaceted, given that primary care and specialty care practices often operate with differing training backgrounds, as well as financial incentives, and therefore their ideas of teams must be somewhat reshaped for programs such as this one to succeed. For example, eye clinic professionals (both ophthalmologists and optometrists) need to be convinced that taking in-person DR screening out of their existing workflows—while decreasing the number of nonurgent or benign referrals to their clinic—sufficiently generates enough visits for patients who need higher-level care (possibly including other eye care needs beyond DR). Similarly, primary care professionals need to be educated about the accuracy of telemedicine DR screening compared with in-person examinations, and the need to be assured that appropriate monitoring of the quality of the digital images will ensure accuracy for triaging the scarce specialty care resources. In turn, medical assistants and licensed vocational nurses need to feel confident in adding a new task to their day and that they have sufficient training, support, and feedback to maintain quality control. This investment in clinician and staff education and buy-in from the beginning allowed the telemedicine DR screening program to be deployed using existing staff without outsourcing or hiring, which represents improved health care value (ie, better quality at the same or reduced cost). In doing so, Daskivich et al1 provide a clear example of how safety net health care systems can be leaders in innovation. Although the mainstream view is that innovation often starts within more wellresourced systems and then is adapted and disseminated to other systems once the evidence base is sufficiently deep, it is not uncommon for safety net settings to be better positioned to develop and implement innovations first. The fragmentation within the US health care system and its often lopsided financial incentives across health care systems can create particularly challenging external drivers against efficiency and team-based care. Although safety net health care systems are rarely fully integrated, their longstanding Author Audio Interview