Publications

BACKGROUND

In 2012, the California Department of Public Health began using pyrosequencing (PSQ) to detect mutations associated with resistance to isoniazid, rifampin, quinolones and injectable drugs in Mycobacterium tuberculosis complex. We evaluated the impact of the PSQ assay on the clinical management of tuberculosis (TB) in California.

METHODS

TB surveillance and laboratory data for specimens submitted 1 August 2012 through 31 December 2016 were analyzed to determine time to effective treatment initiation. A survey of clinicians was used to assess how PSQ results influenced clinical decision making.

RESULTS

Of 1957 specimens tested with PSQ, 52% were sediments and 46% were culture isolates, submitted a median of 8 and 35 days, respectively, after collection. Among 36 patients with multidrug-resistant (MDR) TB who had a sediment specimen submitted for PSQ, median time from specimen collection to MDR-TB treatment initiation was 12 days vs 51 days when PSQ was not used. Completed surveys were returned for 303 patients, 177 of whom reported a treatment change; 75 (42%) of clinicians reported PSQ as a reason for change. Twenty-one patients either had an MDR-TB risk factor and a smear-positive sputum specimen, but had PSQ performed on a culture isolate (9/36 [25%]); or did not have PSQ used for MDR-TB diagnosis (12/38 [32%]) and thus had an opportunity for earlier MDR-TB diagnosis with PSQ on sediment.

CONCLUSIONS

Patients with MDR-TB initiated effective treatment 5 weeks earlier when PSQ was used compared to those without PSQ. Survey data suggest clinicians use PSQ to devise effective TB drug regimens. To maximize the benefit of PSQ, earlier submission of specimens should be prioritized.

Aiming to increase care access, the national Primary Care-Mental Health Integration (PC-MHI) initiative of the Veterans Health Administration (VHA) embedded specialists, care managers, or both in primary care clinics to collaboratively care for veterans with psychiatric illness. The initiative's effects on health care use and cost patterns were examined among 5.4 million primary care patients in 396 VHA clinics in 2013-16. The median rate of patients who saw a PC-MHI provider was 6.3 percent. Each percentage-point increase in the proportion of clinic patients seen by these providers was associated with 11 percent more mental health and 40 percent more primary care visits but also with 9 percent higher average total costs per patient per year. At the mean, 2.5 integrated care visits substituted for each specialty-based mental health visit that did not occur. PC-MHI was associated with improved access to outpatient care, albeit at increased total cost to the VHA. Successful implementation of integrated care necessitates significant investment and multidisciplinary partnership within health systems.

BACKGROUND/OBJECTIVE

Offering depression collaborative care services in primary care (PC) settings can reduce use of nonintegrated mental health care resources and improve mental health care access, particularly for vulnerable PC patients. Tests of effects on depression care quality, however, are needed. We examined overall quality of depression care and tested whether increasing clinic engagement in Veterans Affairs (VA)'s Primary Care-Mental Health Integration (PC-MHI) services was associated with differences in depression care quality over time.

METHODS

We conducted a retrospective longitudinal cohort study of 80,136 Veterans seen in 26 Southern California VA PC clinics (October 1, 2008-September 30, 2013). Using multilevel regression models adjusting for year, clinic, and patient characteristics, we predicted effects of clinic PC-MHI engagement (ie, percent of PC patients receiving PC-MHI services) on 3 VA-developed longitudinal electronic population-based depression quality measures among Veterans newly diagnosed with depression (n=12,533).

RESULTS

Clinic PC-MHI engagement rates were not associated with significant depression care quality differences. Across all clinics, average rates of follow-up within 84 or 180 days were, 66.4% and 74.5%, respectively. Receipt of minimally appropriate treatment was 80.5%. Treatment probabilities were significantly higher for vulnerable PC patients (homeless: 4.5%, P=0.03; serious mental illness: 15.2%, P<0.001), than for otherwise similar patients without these characteristics.

CONCLUSIONS/POLICY IMPLICATIONS

Study patients treated in PC clinics with greater PC-MHI engagement received similarly high quality depression care, and even higher quality for vulnerable patients. Findings support increasing use of PC-MHI models to the extent that they confer some advantage over existing services (eg, access, patient satisfaction) other than quality of care.