There was considerable individual practice variability in these treatment behaviors. For example, although the average incidence for all practices in asking about mood in symptomatic patients was 35%, rates for the different practices ranged from 21% to 65%. Similarly, although an average of 15% of respondents were given a follow-up appointment, this ranged from 6% in one practice to 48% in another.
Additional information from the analysis (not shown in Table 2) showed that 20 (27%) of the 75 patients scoring higher than 1.75 neither mentioned mood themselves nor reported that the physician asked about it, and they were not receiving any type of mental health treatment. Of the 139 total patients the physicians did not ask about mood, 35 (25%) were taking a psychotropic medication (24, or 69%, of these were antidepressants). Of the 71 patients who reported that they were sad, down, or depressed at the visit, only 33 (46%) reported being asked about their mood by the physician. Only 11 (16%) of these 71 patients reported that the physician provided any counseling at the visit.
Discussion
Whether one relies on qualitative interview data, quantitative clinician self-report data, or quantitative patient report data, these results suggest that current care patterns for patients with depression are unsystematic and greatly variable. Although a score on the HSCL-d20 is not the same as a clinical diagnosis of depression, the issue of mood or stress was not brought up by the clinician in two thirds of the visits with patients reporting lower levels of depressive symptoms or in more than half (56%) of those scoring in the range where major depression is a possibility. Less than half (46%) of the patients who considered themselves depressed reported being asked about this by the clinician. Other studies of the accuracy of patient report suggest that patient report is either reasonably similar to physician report or errs on the side of overreporting physician behavior, so this lack of addressing depression is of concern.29-30
The focus groups suggest little clinician support for the time and effort of systems to routinely screen all visiting patients for depression. The scientific literature thus far concurs that there is little evidence that more systematic identification of cases of depression would necessarily be helpful.5,18, 31-32 However, all 3 data sources (focus groups, physician surveys, and patient surveys) suggest that management approaches in these practices have substantial room for improvement. The primary reliance on medications with little attention to the use of other treatment strategies or to any systematic follow-up is particularly troublesome, since recent studies suggest that a wide variety of approaches can add to the efficacy of treatment.2,27,33-34 There are increasing indications that intensive follow-up and relapse prevention is necessary for many patients, supporting the need for more systematic care.13,18,35-36
The clinicians studied not only accepted the responsibility of treating depression, but also seemed open to various ways to improve care. They appeared especially ready to entertain more organized monitoring and follow-up and on-site collaboration with mental health counselors, as long as the issues of care complexity, role clarification, and costs could be worked out. The main barriers they identified in current care—lack of time, unavailability of counselors, and patient reluctance—are all potentially resolvable if support systems could be put in place.
Systems that are similar to those that would be needed to treat depression have been well studied and demonstrated to be effective for delivering clinical preventive services in primary care settings.20,37-40 There is also growing evidence and support for a systems approach to improving chronic care as well.41-42 In theory, systems should be ideal for improving the identification of patients with depression, but our data—as well as the evidence that improving identification without first improving treatment is not helpful—suggest that management and follow-up should be the main targets of improvement efforts.
The Approaches to Change
There are no useful models in typical care settings that use comprehensive systems approaches for treating depression and little evidence for the most effective quality improvement change process. The only randomized clinical trial of a continuous quality improvement (CQI) process to improve depression care failed to demonstrate any benefits from the particular process used in its somewhat unusual settings.43 The principal investigator emphasized the difficulty of curbing longstanding clinical habits and the problems of traditional complex and slow CQI approaches.44
However, there appear to be at least 2 promising approaches to overcoming the problems that Goldberg and others have described for traditional CQI methods: (1) an external consultation approach, in which an external group with reason to help facilitate change in independent practices provides those practices with data on their own care processes, plus system tools and training or consultation; and (2) an internal change process approach, in which a medical practice or group of practices charters a multidisciplinary team that uses more recent CQI techniques to conduct its own change process.