• The Art of Mentoring

    The dictionary defines a mentor as “a trusted counselor or guide” (Merriam-Webster, 2018). We have all been influenced by mentors in our lives—positive role models who have taken a personal interest in us and both inspired and guided us to strive to reach our full potential.


  • Sobering Facts about Sober Living

    Stories of victory over substance abuse are compelling: The high school athlete who turns his life of opiate abuse into a college career, the young mother who reunites with her toddler after overcoming meth abuse, and the famous actor who emerges anew after a rehab stint. 


Association between Quality Measures and Mortality in Individuals with Co-Occurring Mental Health and Substance Use Disorders

Feature Articles

People with mental health or substance use disorders (MH/SUDs) have a higher risk of death when compared to the general population, with the highest rates of mortality being among people with co-occurring MH/SUDs (Dickey, Dembling, Azeni, & Normand, 2004; Muhuri & Gfroerer, 2011; Roerecke & Rehm, 2013; Rosen, Kuhn, Greenbaum, & Drescher, 2008; Walker, McGee, & Druss, 2015). Reducing the premature mortality associated with co-occurring MH/SUDs is an ongoing and important public health concern. It is possible that by providing high-quality health care, health care systems could reduce the premature mortality experienced by individuals with co-occurring MH/SUDs, but it is unknown whether providing high-quality care would result in lower mortality.


Currently, there are no validated measures to examine quality of care and its impact on mortality in this patient population (Dausey, Pincus, & Herrell, 2009). This is an important gap, because if quality of care is associated with decreased mortality, health care systems could use these measures to monitor and improve their performance and have some assurance that doing so would result in improved outcomes.


There are two main types of quality measures: process and outcome measures. Process measures evaluate the care provided to patients. Outcome measures evaluate the impact of the care on patients’ symptoms, functional status, mortality or quality of life. While outcome measures are the gold standard for evaluating whether a treatment is beneficial, measuring outcomes can be difficult. Outcome measures are harder to collect and cannot be used to identify where quality falls short and where quality improvement efforts should be targeted. Outcome measures must also be risk adjusted for illness severity, which requires the collection of comorbid conditions. Process measures have three main advantages over outcome measures. First, process measures can be readily obtained from administrative data. Second, process measures give insight into which processes of the health care system need to be improved. Third, they provide this insight in real-time, allowing health care systems the opportunity to take timely corrective action. 


The study described herein describes the association between five potential process-related quality measures and one- and two-year mortality among patients with co-occurring MH/SUDs (Watkins et al., 2016). If these five potential process measures are associated with lower mortality, then there are two important implications. First, an association with mortality provides initial evidence for the validity of these five process measures. Second, it suggests that by improving performance on these measures, health care systems could take concrete actions to decrease the risk of death among this vulnerable population. 


Study Population 


Veterans with either schizophrenia, bipolar I disorder, major depression or posttraumatic stress disorder (PTSD) who also had SUDs were included in these analyses if they had received care from or if their care was paid for by the Veterans Health Administration (VHA) in the fiscal year 2007. Veterans also had to have at least one inpatient hospitalization or two outpatient clinic visits to show they were routinely receiving care from the VHA. 


Quality Measures


The five process measures tested in the current study were (Harris et al., 2015; Watkins et al., 2011): 


  1. Receipt of at least one psychotherapy visit for either a MH/SUD
  2. Receipt of at least one psychosocial treatment visit for either a MH/SUD
  3. Receipt of at least one visit to any type of provider for a MH/SUD per quarter
  4. Treatment initiation
  5. Treatment engagement


Psychotherapy included at least one group or individual psychotherapy visit, and psychosocial treatment visit included at least one individual and group psychotherapy, family intervention, supported employment, skills training or intensive care management visit. Receipt of at least one diagnosis-related visit per quarter was defined as a visit to any provider for either a MH/SUD. Treatment initiation was defined as having received at least one treatment visit for the SUD within fourteen days of the initial diagnosis. Treatment engagement was defined as receipt of at least two treatment visits related to the SUD in the thirty-day period after treatment initiation. Treatment initiation and engagement only applied to those patients beginning a new treatment episode and engagement was also only applied to those that had treatment initiation. A new treatment episode was defined as having no SUD-related visits in the five months prior to the index visit. 


Statistical Analyses


We examined descriptive statistics for twelve- and twenty-four-month mortality outcomes, patient risk-adjustment characteristics, and for the quality measures. We assessed the strength of association between a quality measure and mortality by examining the probability (odds ratio) of mortality at both twelve and twenty-four months for the quality measure and its 95 percent confidence interval. Probabilities of mortality were adjusted based on demographic and clinical characteristics, service-connected disability (i.e., a marker of illness severity and priority access to VA services), and physical health comorbidities. We also estimated the additional effect on mortality of receipt of care measured by the quality measure, holding all other patient characteristics constant, and report on the number of deaths that potentially could have been averted had patients received the respective quality measure.


Additional Analyses


We also looked at the relationship between mortality and number of diagnosis-related outpatient visits. We used the following categories for number of visits: 


  • One to two (reference)
  • Five to ten
  • Eleven to twenty
  • Twenty-one to fifty
  • Fifty-one and over


Because people who are sicker may be more likely to get more care, it is possible that getting more care could be associated with mortality independent of the quality of care provided (Lin, Psaty, & Kronmal, 1998). Receipt of care can be dependent on the severity of patients’ illnesses, which may not be able to be measured based on risk factors found in the data. To evaluate this unmeasurable risk, techniques were employed that would estimate how much influence unmeasured illness severity would have to have in order to change the interpretation of the results.


Association of Quality of Care with Mortality


There were 144,045 veterans included in our analysis that had co-occurring MH/SUDs who used VHA services in fiscal year 2007. About 95 percent of those were male and the average age was fifty-two years old. Around 50 percent of the veterans had a diagnosis for PTSD, and approximately 21 percent had a diagnosis of major depression. Among those with substance use disorders, alcohol abuse or dependence was the most common disorder. Roughly 75 percent of these patients had a minimum of one new treatment episode. These episodes could have been for either a mental health disorder, a substance use disorder or both. 


The mortality rates at twelve and twenty-four months were 2.7 percent and 5.3 percent respectively. The range for twelve-month mortality rates was 2.6 percent for those veterans with co-occurring bipolar disorder to 3 percent for those veterans with co-occurring schizophrenia. The range for twenty-four-month mortality rates were 5.1 percent for veterans with co-occurring PTSD and 5.9 percent for veterans with co-occurring bipolar disorder.
Adherence to the five predefined quality measures was high. Ninety percent of the population received some psychosocial treatment, and 67 percent received psychotherapy. Treatment initiation was 19.8 percent among those veterans with a new treatment episode, and 60.2 percent of those who initiated, engaged with treatment. Forty-one percent had at least one diagnosis-related outpatient visit in each quarter.


Each measure was significantly associated with lower twelve- and twenty-four-month mortality rates except for treatment engagement, which was associated with decreased mortality at twelve months, but not twenty-four months. Across the quality measures, the range of avoidable excess mortality ranged from 19 percent to 31 percent for twelve-month and 9 percent to 22 percent for twenty-four-month mortality. This means that people who received high-quality care were less likely to die than people who did not get the higher quality care. For each quality measure, receiving the higher quality care could have averted up to 984 deaths. The sensitivity analyses suggested that these results were unlikely to be because of unmeasured illness severity.


The association between number of diagnosis-related outpatient visits and twelve-month mortality was also evaluated. A linear and inverse relationship exists between number of outpatient visits and twelve-month mortality. That is, the higher the number of outpatient visits, the lower the risk of twelve-month mortality. This relationship leveled off when the number of visits reached twenty-one to fifty in a twelve-month period. 


Limitations of this Study


Potential limitations exist with this work. First, we cannot be certain whether our results would be generalizable to populations outside of the VHA system. However, since our findings are consistent across various types of visits and measures, it suggests that our findings would likely be valid for patients treated in non-VHA settings. Second, the observational data used for this study are only able to determine associations and not causal mechanisms for changes. This means that while higher quality care was associated with decreased mortality risk, we do not know whether it caused the lower risk of death. Third, this study uses data from fiscal year 2007. It is possible that today’s treatment processes may be different. However, since the association between the five process quality measures were not dependent on type of visit, it is unlikely that the association would be different with time.


Implications for Health Care Systems and Practitioners


Better adherence to the five process quality measures defined in this study was associated with decreased twelve-month and twenty-four-month mortality among the patients with co-occurring MH/SUDs. While this study does not address what type of care or type of provider is optimal, there exists a linear and robust association between number of visits and decreased mortality, suggesting that more visits is better. While one diagnosis-related visit per quarter may be a reasonable standard for health care systems to maintain for patients with co-occurring MH/SUDs, our additional analyses suggest that more visits up to once a week could be associated with even lower risks of death.


This study did not evaluate how improved quality is associated with decreased mortality, although there are several possibilities. The lower risk of death could be from the earlier detection and management of physical comorbidities, earlier recognition of mental health or substance use relapse or engagement in preventive health services (Copeland et al., 2009; Druss, Bradford, Rosenheck, Radford, & Krumholz, 2001; Hayes et al., 2015; Roerecke, Gual, & Rehm, 2013). Previous research has found that veterans with mental illness who dropped out of care for prolonged periods experienced a six-fold reduction in mortality after reengagement and utilization of VA services as compared to those veterans who did not return to the VA for care (Davis et al., 2012). Treatment could also result in less alcohol and other drug use. Our results are in line with previous research that shows early follow-up for patients with SUDs after residential treatment is associated with decreased two-year mortality (Harris et al., 2015).


Mental illness and serious mental illness shortens the average lifespan by an average of 8.2 and eleven to twenty-five years respectively, as compared to the general population (Colton & Manderscheid, 2006; Druss, Zhao, Von Esenwein, Morrato, & Marcus, 2011). SUDs are also associated with premature mortality (Dickey et al., 2004; Rosen et al., 2008; Yoon, Chen, Yi, & Moss, 2011), with the highest association being within those patients with co-occurring MH/SUDs (Dickey et al., 2004; Maynard, Cox, Hall, Krupski, & Stark, 2004). The measure of service utilization used for this study was at least one visit related to either mental illness or substance use per quarter with any type of provider. This points to the importance of all providers being aware of the presence of these diagnoses and the collaborative effort that all providers play in reducing mortality. 


These five quality measures can all be operationalized using readily available administrative data, and many of the EHR systems on the market should be able to be tailored to report on these measures, giving an opportunity for real-time monitoring and response by the health care system. Using administrative data from their EHRs, health care systems and clinics should consider monitoring the following quality measures for patients with co-occurring MH/SUDs:


  • Ensure at least one additional substance use treatment visit within fourteen days of a new SUD diagnosis.
  • Ensure at least two additional SUD-related treatment visits within the month after the initiation visit
  • Offer at least one diagnosis-related visit for either the MH/SUD with any provider type per quarter, with up to one to two visits per month if possible
  • Ensure that both psychotherapy and psychosocial treatments are available to all patients with co-occurring MH/SUD and encourage patients to make use of them










Colton, C. W., & Manderscheid, R. W. (2006). Congruencies in increased mortality rates, years of potential life lost, and causes of death among public mental health clients in eight states. Preventing Chronic Disease, 3(2), A42.
Copeland, L. A., Zeber, J. E., Wang, C. P., Parchman, M. L., Lawrence, V. A., Valenstein, M., & Miller, A. L. (2009). Patterns of primary care and mortality among patients with schizophrenia or diabetes: A cluster analysis approach to the retrospective study of healthcare utilization. BMC Health Services Research, 9, 127.
Dausey, D. J., Pincus, H. A., & Herrell, J. M. (2009). Performance measurement for co-occurring mental health and substance use disorders. Substance Abuse Treatment, Prevention, and Policy, 4, 18.
Davis, C. L., Kilbourne, A. M., Blow, F. C., Pierce, J. R., Winkel, B. M., Huycke, E., . . . Visnic, S. (2012). Reduced mortality among Department of Veterans Affairs patients with schizophrenia or bipolar disorder lost to follow-up and engaged in active outreach to return for care. American Journal of Public Health, 102(Suppl. 1), S74–9.
Dickey, B., Dembling, B., Azeni, H., & Normand, S. L. (2004). Externally caused deaths for adults with substance use and mental disorders. Journal of Behavioral Health Services and Research, 31(1), 75–85.
Druss, B. G., Bradford, W. D., Rosenheck, R. A., Radford, M. J., & Krumholz, H. M. (2001). Quality of medical care and excess mortality in older patients with mental disorders. Archives of General Psychiatry, 58(6), 565–72.
Druss, B. G., Zhao, L., Von Esenwein, S., Morrato, E. H., & Marcus, S. C. (2011). Understanding excess mortality in persons with mental illness: Seventeen-year follow-up of a nationally representative US survey. Medical Care, 49(6), 599–604.
Harris, A. H., Gupta, S., Bowe, T., Ellerbe, L. S., Phelps, T. E., Rubinsky, A. D., . . . Trafton, J. (2015). Predictive validity of two process-of-care quality measures for residential substance use disorder treatment. Addiction Science & Clinical Practice, 10, 22.
Hayes, R. D., Downs, J., Chang, C. K., Jackson, R. G., Shetty, H., Broadbent, M., . . . Stewart, R. (2015). The effect of clozapine on premature mortality: An assessment of clinical monitoring and other potential confounders. Schizophrenia Bulletin, 41(3), 644–55.
Lin, D. Y., Psaty, B. M., & Kronmal, R. A. (1998). Assessing the sensitivity of regression results to unmeasured confounders in observational studies. Biometrics, 54(3), 948–63.
Maynard, C., Cox, G. B., Hall, J., Krupski, A., & Stark, K. D. (2004). Substance use and five-year survival in Washington State mental hospitals. Administration and Policy in Mental Health and Mental Health Services Research, 31(4), 339–45.
Muhuri, P. K., & Gfroerer, J. C. (2011). Mortality associated with illegal drug use among adults in the United States. American Journal of Drug and Alcohol Abuse, 37(3), 155–64.
Roerecke, M., Gual, A., & Rehm, J. (2013). Reduction of alcohol consumption and subsequent mortality in alcohol use disorders: Systematic review and meta-analyses. Journal of Clinical Psychiatry, 74(12), e1181–9.
Roerecke, M., & Rehm, J. (2013). Alcohol use disorders and mortality: A systematic review and meta‐analysis. Addiction, 108(9), 1562–78.
Rosen, C. S., Kuhn, E., Greenbaum, M. A., & Drescher, K. D. (2008). Substance abuse-related mortality among middle-aged male VA psychiatric patients. Psychiatric Services, 59(3), 290–6.
Walker, E. R., McGee, R. E., & Druss, B. G. (2015). Mortality in mental disorders and global disease burden implications: A systematic review and meta-analysis. JAMA Psychiatry, 72(4), 334–41.
Watkins, K. E., Paddock, S. M., Hudson, T. J., Ounpraseuth, S., Schrader, A. M., Hepner, K. A., & Sullivan, G. (2016). Association between quality measures and mortality in individuals with co-occurring mental health and substance use disorders. Journal of Substance Abuse Treatment, 69, 1–8.

Watkins, K. E., Smith, B., Paddock, S. M., Mannle Jr., T. E., Woodroffe, A., Solomon, J., . . . Pincus, H. A. (2011). Veterans Health Administration mental health program evaluation: Capstone report. Retrieved from https://www.mentalhealth.va.gov/docs/capstone_revised_tr956_compiled.pdf

Yoon, Y. H., Chen, C. M., Yi, H. Y., & Moss, H. B. (2011). Effect of comorbid alcohol and drug use disorders on premature death among unipolar and bipolar disorder decedents in the United States, 1999 to 2006. Comprehensive Psychiatry, 52(5), 453–64.
Editor’s Note: This article was adapted from an article by the same authors previously published in the Journal of Substance Abuse Treatment (JSAT). This article has been adapted as part of Counselor’s memorandum of agreement with JSAT. The following citation provides the original source of the article:
Watkins, K. E., Paddock, S. M., Hudson, T. J., Ounpraseuth, S., Schrader, A. M., Hepner, K. A., & Sullivan, G. (2016). Association between quality measures and mortality in individuals with co-occurring mental health and substance use disorders. Journal of Substance Abuse Treatment, 69, 1–8.