http://orcid.org/0000-0001-8862-8078
Abstract
Background: Medical comorbidity may influence treatment initiation and engagement for alcohol and other drug (AOD) use disorders. We examined the association between medical comorbidity and Healthcare Effectiveness Data and Information Set (HEDIS) treatment initiation and engagement measures.
Methods: We used electronic health record and insurance claims data from 7 US health care systems to identify patients with AOD use disorders between October 1, 2014, and August 15, 2015 (N = 86,565). Among patients identified with AOD use disorders in outpatient and emergency department (ED) settings, we examined how Charlson/Deyo comorbidity index scores and medical complications of AOD use were associated with treatment initiation. Among those who initiated treatment in inpatient and outpatient/ED settings, we also examined how comorbidity and AOD use–related medical complications were associated with treatment engagement. Analyses were conducted using generalized estimating equation logistic regression modeling.
Results: Among patients identified as having an AOD diagnosis in outpatient and ED settings (n = 69,965), Charlson/Deyo comorbidity index scores of 2 or more were independently associated with reduced likelihood of initiation (risk ratio [RR] = 0.80, 95% confidence interval [CI] = 0.74, 0.86; reference score = 0), whereas prior-year diagnoses of cirrhosis (RR = 1.25, 95% CI = 1.12, 1.35) and pancreatic disease (RR = 1.34, 95% CI = 1.15, 1.56) were associated with greater likelihood of initiation. Among those who were identified in outpatient/ED settings and initiated, higher comorbidity scores were associated with lower likelihood of engagement (score 1: RR = 0.85, 95% CI = 0.76, 0.94; score 2+: RR = 0.61, 95% CI = 0.53, 0.71).
Conclusion: Medical comorbidity was associated with lower likelihood of initiating or engaging in AOD treatment, but cirrhosis and pancreatic disease were associated with greater likelihood of initiation. Interventions to improve AOD treatment initiation and engagement for patients with comorbidities are needed, such as integrating medical and AOD treatment.
Introduction
Healthcare Effectiveness Data and Information Set (HEDIS) measures are influential performance measures for health insurance plans and delivery systems because they are used to generate health plan rankings and may be linked to financial incentives. As a result, such measures can help promote and guide quality improvement initiatives. For alcohol and other drug (AOD) use disorders, HEDIS measures were developed by the Washington Circle and implemented by the National Committee for Quality Assurance (NCQA)Citation1 and the Department of Veterans Affairs.Citation2 These measures are used to measure initiation and engagement rates in treatment for AOD use disorders.
In community and specialty treatment settings, several factors have been associated with AOD treatment initiation, including sociodemographic factors (e.g., age,Citation3 race,Citation4–6 sex,Citation3,Citation7 employmentCitation7), addiction severity,Citation7 recent arrest,Citation8 and psychiatric comorbidity.Citation3,Citation8 Additionally, the setting of care in which the individual is identified with an AOD use disorder (e.g., inpatient medical, inpatient AOD specialty, and outpatient AOD specialty treatment) has been strongly associated with treatment initiation and engagement.Citation9 In particular, engagement has been shown to be higher for individuals identified in specialty AOD treatment settings compared with general medical settings.Citation9 Patients with AOD use disorders who are identified upon entering AOD treatment may differ from those who are identified in medical settings in a myriad of ways that impact treatment initiation and engagement rates, including their readiness and ability to access treatment and the severity of their medical conditions. Yet little is currently known about how medical comorbidity impacts AOD treatment initiation and engagement.
Medical problems may be associated with AOD initiation and engagement for several reasons. First, medical problems that result from or are exacerbated by substance use, such as human immunodeficiency virus (HIV), hepatitis C, pancreatitis, or overdose, are often markers of AOD use disorder severity.Citation10 As such, they may be associated with improved AOD treatment initiation and engagement because they could motivate physicians to identify AOD use disorders, refer patients to AOD treatment, and facilitate engagement.Citation11 These medical conditions may also motivate patients to enter or remain in treatment and encourage family members to support patient initiation and engagement. Some providers require patients to complete AOD treatment as a precondition of receiving treatment for co-occurring health conditions, such as hepatitis C or liver transplants.Citation12 Even for medical comorbidities that are largely independent of substance use (e.g., chronic renal insufficiency, diabetes), frequent contact with the health care system could increase opportunities for screening, diagnosis, and referral to AOD treatment.
On the other hand, medical comorbidities may also negatively influence individuals’ ability to initiate and engage in AOD treatment.Citation13 Medical comorbidities may have cognitive effects, physical symptoms, and associated disabilities that make it difficult for patients to attend the required number of visits within the time windows specified by the HEDIS measures. Further, they may lead to competing demands and higher health care costs that limit participation in AOD treatment. For example, conditions such as congestive heart failure may lead to a high baseline burden of medical visits, medical costs, and symptoms, such as dyspnea on exertion. Further, if medical complications of AOD use act as a marker of AOD use disorder severity, then attendance in the requisite visits to meet engagement criteria (2 within 30 days of initiation or inpatient discharge) may be more challenging due to the AOD use disorder itself.
Given these competing hypotheses, we sought to determine the independent associations between a global measure of medical comorbidity and specific medical complications of substance use with HEDIS-defined AOD treatment initiation and engagement. Gaining a better understanding of the impact of medical comorbidity on the HEDIS initiation and engagement measures can inform interventions to improve access and delivery of AOD treatment services to medically complex patients in large health systems. Further, initiation and engagement rates provide benchmarks to determine progress on achieving recommended practices.
Methods
Study design, settings, and population
This multisite retrospective cohort study was based on data from 7 US health care delivery systems that are members of the Health Systems Node of the National Institute on Drug Abuse (NIDA) Clinical Trials Network. We examined AOD treatment initiation and engagement between October 1, 2014, and August 15, 2015, among adult patients (age ≥18) who qualified for the HEDIS measure denominatorCitation14,Citation15 and were continuously enrolled in one of the health systems 60 days prior to through 44 days after the index date. This research was reviewed and approved by the Kaiser Permanente Northern California Institutional Review Board. This study met requirements for a waiver of informed consent.
Data sources
Data were extracted from each site’s Virtual Data Warehouse (VDW), a standardized data model that was developed for research use across the Health Care Systems Research Network (HCSRN).Citation16,Citation17 The VDW represents a common data structure for electronic health record (EHR), insurance claims, and other administrative data for research purposes that have been extracted and loaded into relational tables linked through a common, unique identifier.Citation18–20 VDW diagnosis and procedure files include coded diagnoses and procedures associated with inpatient and outpatient encounters. These codes are based on International Classification of Diseases, Ninth edition, Clinical Modification (ICD-9-CM), Healthcare Common Procedure Coding System (HCPCS), and the Fourth Edition of the Common Procedure Terminology codes (CPT-4).
Outcomes—initiation and engagement
Data elements required for calculating the HEDIS performance measures were extracted from the VDW databases and included diagnosis-related group categories, ICD-9-CM diagnosis codes, CPT codes, Uniform/Universal Billing Form 92 Revenue codes, Centers for Medicare & Medicaid Services (CMS) 1500 site of service codes, department, and date of services. Following the NCQA Measure Technical Specifications, AOD initiation and engagement rates were calculated only for adult patients who had a “new” index episode (defined by having a period of at least 60 days before the episode without a diagnosis of AOD abuse or dependence) with an AOD abuse or dependence diagnosis between October 1, 2014, and August 15, 2015.Citation14 An index identification setting could be an outpatient, observation/emergency department (ED), or inpatient claim/encounter/discharge, including AOD specialty treatment. If more than 1 diagnosis qualified an individual for inclusion, the first (primary diagnosis) classified the index as alcohol versus drug abuse/dependence. If the index episode was any inpatient discharge, consistent with the HEDIS definition of initiation, the inpatient stay was considered to have met treatment initiation, regardless of whether the admission was to an inpatient AOD treatment program or a medical hospitalization. If the index episode was an ED or outpatient claim/encounter, the patient had to have a subsequent AOD service (not including an ED visit or detoxification inpatient stay) within 14 days of the index episode to be considered “initiated.” For the engagement measure, patients who had 2 or more AOD-related services within 30 days after initiating treatment (or hospital discharge) were considered “engaged.”
Study variables
We examined patient characteristics, including age, sex, and race/ethnicity. We extracted all ICD-9-CM AOD abuse/dependence diagnoses in the year prior to the index identification visit, including alcohol, opioid, barbiturate, cocaine, cannabis, amphetamine, hallucinogen, and unspecified. Based on diagnosis codes in the year prior to the index date, the Deyo version of the Charlson comorbidity index was calculated for all patients.Citation21,Citation22 The Charlson/Deyo index was designed to account for comorbidity and disease severity using ICD-9 and procedure codes, with increasing score associated with worse outcomes, including 1-year mortality. Conditions were given weighted scores, ranging from 1 to 6. For our study, we categorized scores as 0, 1, and 2 or more, based on the distribution of scores in our sample. For medical complications of AOD use or other AOD-related conditions, we extracted medical and psychiatric diagnoses in the year prior to the index identification date based on the Healthcare Cost and Utilization Project (HCUP) classification system.Citation23 We also extracted 21 substance abuse–related medical conditions (SAMCs) based on the ICD-9-CM diagnosis codes.Citation24 From the HCUP and SAMC categories, we examined AOD diagnostic groups based on clinical relevance for our research questions. Although not a medical complication of AOD use per se, we included pregnancy and childbirth due to the potential for poor outcomes associated with untreated AOD use disorders.
Statistical analysis
Our analyses examined the association between medical comorbidity and the following: (1) treatment initiation among patients identified with an AOD use disorder in outpatient and ED settings; (2) treatment engagement among patients identified with an AOD use disorder in any inpatient setting (includes AOD use disorder and medical treatment settings) and initiated treatment; and (3) treatment engagement among patients identified with an AOD use disorder in outpatient and ED settings (includes outpatient AOD use disorder treatment settings) and initiated treatment. There are strong associations between setting and initiation/engagement rates,Citation9 and HEDIS criteria for initiation are met if the index encounter was for inpatient treatment. Therefore, we could not examine factors associated with treatment initiation in the inpatient setting and conducted analyses for initiation only among patients identified in the outpatient and ED settings. We checked data quality, consistency, and correlations prior to statistical modeling. If 2 covariates were highly correlated (r > .40), only one of the variables was included in the model, based on an assessment of clinical relevance. We described patient characteristics using percentages or medians with interquartile ranges for all variables except age (which was described by a mean with standard deviation) and Charlson/Deyo (which was also described by a median with interquartile ranges). For descriptive purposes, differences between those who initiated or engaged in treatment and those who did not were examined with Wilcoxon rank-sum tests and chi-square tests of association for interval-level and categorical variables, respectively. For our primary analyses, risk ratios (RRs) were estimated from a generalized estimating equation logistic regression model using health plan as a clustering variable. All variables with a P < .25 in the bivariate analyses were considered for inclusion in the final adjusted models. Sex was included in all models due to its strong association with comorbidity. Forward selection was used to identify other variables significantly (P < .05) associated with initiation and engagement. Variables in the final models were assessed for multicollinearity. All analyses were performed using SAS 9.4 (SAS Insitute, Cary, NC). In supplemental analyses, we also conducted a combined analysis examining factors associated with treatment engagement among individuals who met initiation criteria in any setting.
Results
During the study period, 86,565 patients with an index AOD use disorder diagnosis were eligible for inclusion in the cohort across the 7 health plans/systems. Across all index identification settings, the median Charlson/Deyo comorbidity index score was 0 (interquartile range [IQR] = 0, 1), with a range of 0–20 (). Nearly one third (30.6%, n = 10,276) of patients with any comorbidity received their index diagnosis of an AOD disorder in an inpatient setting.
Table 1. Demographic and clinical characteristics of a study cohort of patients with alcohol and other drug (AOD) use disorders at 7 health systems (N = 86,565) stratified by AOD treatment initiation and engagement status, October 2014 through August 2015.
describes the prior-year prevalence of conditions across HCUP and SAMC diagnoses for all eligible patients. Mental disorders, including AOD abuse/dependence, was the most prevalent specified past-year HCUP category (66.2%), followed by musculoskeletal and connective tissue disorders (52.1%) and nervous system/sense organ conditions (50.8%). Injury/poisoning was the most prevalent past-year SAMC condition (35.5%), followed by hypertension (34.8%) and depression (30.4%).
Table 2. Among all patients with an alcohol or drug (AOD) in 7 health systems, prior-year prevalence of Healthcare Cost and Utilization Project (HCUP)Citation23 body system disease or disorder and any of 21 substance abuse-related medical conditions (SAMCs).Citation24
Initiation of treatment
We identified treatment initiation in 27.9% (n = 24,188) of eligible patients, of whom approximately half had an index diagnosis of alcohol use disorder (50.7%, n = 12,252); the remainder had an index diagnoses of drug use disorder (49.3%; n = 11,936; ). Across all settings, those who initiated treatment were significantly older (median age 52.0 years, IQR = 35.0, 64.0) than those who did not (47.0 years, IQR = 31.0, 60.0). They were also more likely to have a diagnosis of tobacco dependence (29.5% vs. 19.5%), pregnancy/childbirth (4.2% vs. 1.5%), and 1 or more medical complications of substance use (injury/poisoning, cirrhosis, hepatitis C, or pancreatic disease; 40.2% vs. 24.9%) in the prior year compared with those who did not initiate. A higher proportion of those who initiated had Charlson/Deyo comorbidity index scores of 2 or more (33.5%) compared with those who did not initiate (17.0%).
Among patients who were identified as having an AOD diagnosis in outpatient and ED settings (n = 69,965; ), Charlson/Deyo comorbidity index scores of 2 or more were independently associated with a reduced likelihood of initiation (score 1: RR = 0.99, 95% confidence interval [CI] = 0.96, 1.03; score 2+: RR = 0.80, 95% CI = 0.74, 0.86; reference score = 0). Patients with a prior-year diagnosis of tobacco dependence (RR = 1.30, 95% CI = 1.17, 1.43) or alcohol abuse (RR = 1.57, 95% CI = 1.32, 1.86) were more likely to initiate treatment compared with those without these diagnoses. Further, prior-year diagnoses of cirrhosis (RR = 1.25, 95% CI = 1.12, 1.35) and pancreatic disease (RR = 1.34, 95% CI = 1.15, 1.56) were associated with higher likelihood of initiation.
Table 3. Among those identified with an alcohol or drug (AOD) use disorder in an outpatient or emergency department setting (n = 69,965), characteristics associated with initiation of AOD treatment.
Treatment engagement
Among those who were diagnosed with AOD use disorder in any setting (outpatient, ED, or inpatient) and initiated treatment, 11.5% (n = 2,782) met HEDIS criteria for AOD treatment engagement (). Patients who engaged in AOD treatment were younger (44.0 years, IQR = 30.0, 55.0) and more likely to be non-Hispanic white (64.2%) compared with those who did not engage (median age 53.0 years, IQR = 36.0, 65.0; 62.1% non-Hispanic white). For those who engaged, the index visit type was more likely to be an outpatient visit (61.6%) compared with those who did not engage (16.7%). Those who engaged also had less medical comorbidity (Charlson/Deyo index score of 2+: 16.1% vs. 35.8%) and fewer medical complications of substance use (injury/poisoning, liver cirrhosis, hepatitis C, or pancreatic disease; 43.2% vs. 55.5%).
Among patients who were identified as having an AOD diagnosis in an inpatient setting (n = 16,193; ), Charlson/Deyo comorbidity scores of 2 or more were less likely to engage (RR = 0.70, 95% CI = 0.61, 0.79) compared with those with scores of 0. Factors associated with increased engagement included past-year alcohol dependence (RR = 2.70, 95% CI = 2.23, 3.26), depressive disorder (RR = 1.37, 95% CI = 1.23, 1.52), and major psychotic disorder (RR = 1.47, 95% CI = 1.15, 1.88).
Table 4. Among patients identified with an alcohol or drug use disorder (AOD) in an inpatient setting who initiated AOD treatment (n = 16,193), characteristics associated with AOD treatment engagement.
When we examined treatment engagement only in the subset of patients who were identified in the outpatient or ED settings and subsequently initiated treatment (n = 7,963; ), factors associated with decreased likelihood of engagement included Charlson/Deyo comorbidity scores of 1 (RR = 0.85, CI = 0.76, 0.94) or 2 or more (RR = 0.61, CI = 0.53, 0.71, compared with those with scores of 0), being African American (RR = 0.76, 95% CI = 0.66, 0.87), and index identification in the ED setting (RR = 0.47, 95% CI = 0.42, 0.52).
Table 5. Among patients identified as having an alcohol or drug (AOD) use disorder in an outpatient or emergency department setting and initiated treatment (n = 7,963), characteristics associated with AOD treatment engagement.
Among patients who were identified as having an AOD diagnosis in any setting and initiated treatment (n = 24,188; ), Charlson/Deyo comorbidity scores of 1 (RR = 0.87, 95% CI = 0.82, 0.93) or 2 or more (RR = 0.65, 95% CI = 0.58, 0.72) were less likely to engage compared with those with scores of 0. Factors associated with increased engagement included female sex (RR = 1.02, 95% CI = 1.00, 1.05), index encounter types of outpatient, ED, or other (RR = 6.73, 95% CI = 5.07, 8.92; RR = 3.31, 95% CI = 2.45, 4.46; and RR = 3.40, 95% CI = 1.68, 6.89, respectively, compared with inpatient index types), and a past-year diagnosis of pancreatic disease (RR = 1.23, 95% CI = 1.20, 1.26). In contrast, factors associated with reduced engagement included being African American (RR = 0.73, 95% CI = 0.63, 0.84, compared with non-Hispanic white race/ethnicity), pregnancy/childbirth (RR = 0.62, 95% CI = 0.45, 0.86), and past-year injury or poisoning (RR = 0.97, 95% CI = 0.96, 0.99).
Table 6. Among patients identified with an alcohol and other drug (AOD) use disorder in any setting who initiated AOD treatment (n = 24,188), characteristics associated with AOD use disorder treatment engagement.
Discussion
In this large cohort of health system patients with AOD use disorders, significant general comorbidity was independently associated with poorer adherence to HEDIS AOD treatment initiation and engagement measures. Individuals with greater comorbidity had lower rates of initiation and engagement than patients with no identified medical comorbidity (RR estimates from 0.61 to 0.90). The only medical conditions associated with higher rates of initiation and/or engagement were cirrhosis (initiation only) and pancreatic disease (initiation and engagement).
There are several potential explanations for lower initiation and engagement among individuals with comorbidities. Patients with chronic illness may be more likely to face barriers to attending treatment visits, such as transportation problems, functional disability, or pain. Therefore, they may be less likely to meet HEDIS initiation and engagement criteria. Further quantitative and qualitative research could help identify barriers to receipt of AOD treatment for individuals with comorbidities.
To address barriers to initiation and engagement, telephone or video visits or evidence-based online treatment options, not currently included in the HEDIS AOD measures, may be more convenient for individuals with significant comorbidity.Citation25 Additionally, consistent access to and better integration of AOD treatment and medical services may be needed.Citation26 One approach is to provide AOD treatment at sites where patients obtain primary and specialty medical care.Citation24,Citation27 Analogous to the “medical home” model often used for HIV patient care,Citation28 addiction-trained behavioral health personnel for individual and group visits can be co-located in primary care clinics (i.e., collaborative care), and primary care providers and medical specialists can be trained to provide pharmacologic treatment for AOD use disorders, such as naltrexone and buprenorphine.Citation29 Conversely, physician, nursing, laboratory, or primary care services could be provided in AOD treatment settings by family medicine or internal medicine physicians. On-site medical care at AOD treatment settings has been associated with greater receipt of medical services and fewer ED visits and hospitalizations.Citation30–32
Another factor that may have contributed to the finding that comorbidity was negatively associated with treatment initiation is if patients reduced their AOD use without seeking formal treatment services. For example, there is evidence indicating that as individuals get older and/or develop chronic medical problems, they cut back or stop consuming alcohol.Citation33–35 Alcohol use disorder was the most common AOD use disorder diagnosis in the sample, which included AOD use disorders across the range of severity. Therefore, it is possible that some individuals with significant health problems were motivated to reduce AOD use without initiating treatment, particularly with increasing implementation of screening and brief interventions in health systems.
It is possible that HEDIS measures at the time our study was conducted were inappropriate for patients with substantial comorbidity. For example, some individuals may have met initiation criteria just because they were hospitalized for medical reasons without receiving AOD use disorder treatment in the inpatient setting. It may have been appropriate for the patient to prioritize competing medical needs due to reasons of medical acuity or for financial reasons. However, the HEDIS measures remain valuable in helping to identify gaps in initiation and engagement for patients identified in outpatient and ED settings.
Our results suggest that certain medical conditions, such as cirrhosis and pancreatic disease, may be markers of AOD use disorder severity prompting providers to diagnose and refer patients to treatment. Further research could help determine whether providers are more effective in discussing the need for treatment in the context of certain medical conditions (e.g., pancreatitis) in contrast to other conditions (e.g., overdose or hepatitis C). Such differences could account for our variable results across conditions.
Our findings were consistent with prior studies demonstrating that health plan performance on HEDIS AOD measures is suboptimal and provide additional evidence of the importance of setting of diagnosis on adherence to HEDIS measures, as well as racial and ethnic differences in initiation.Citation6,Citation9 Our findings regarding somewhat lower engagement among patients with pregnancy or childbirth in the prior year are concerning due to the risk of poor outcomes among neonates and the risk of suicide and unintentional overdose among women in the postpartum period.Citation36 These findings warrant further investigation.
Our study had several limitations. We did not examine the reasons for inpatient admission and the type of treatment (i.e., medical, psychiatric, or AOD treatment) in inpatient settings. Further, the widely used performance measures we examined do not include other important dimensions of quality AOD treatment, such as receipt of medication-assisted treatment (e.g., office-based treatment with buprenorphine or naltrexone), the type of behavioral treatment provided, and the patient-centeredness of AOD use disorder treatment services. These may warrant incorporation into future HEDIS metrics.
In this study using data from multiple health plans across the United States, we found that medical comorbidity was associated with poor AOD treatment initiation and engagement. These findings help to highlight several important directions for future research and quality improvement within health systems. First, large-scale health system data could be used to examine the association between meeting AOD performance metrics and medical outcomes such as hospitalization and mortality. Such findings could help refine quality AOD treatment metrics for patients with comorbidity. Second, 2018 HEDIS initiation and engagement measures include medication-assisted treatment and telehealth,Citation37 which may improve initiation and engagement rates for individuals with comorbidity. Further quantitative and qualitative research can identify barriers to initiation and engagement at the patient, provider, health system, and policy levels. Finally, interventions and programs to enhance initiation and engagement for individuals with medical comorbidity should be developed and tested.
Acknowledgments
We thank Morgan Ford, MS, LeeAnn Quintana, MSW, and Jeff Holzman, BS, at Kaiser Permanente Colorado, Andrea Altschuler at Kaiser Permanente Northern California, and Richard Contreras at Kaiser Permanente Southern California for their contributions.
Additional information
Funding
Notes on contributors
Ingrid A. Binswanger
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Nikki M. Carroll
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Brian K. Ahmedani
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Cynthia I. Campbell
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Irina V. Haller
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Rulin C. Hechter
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Jennifer McNeely
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Bobbi Jo H. Yarborough
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Andrea H. Kline-Simon
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Derek D. Satre
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Constance Weisner
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
Gwen T. Lapham
I.A.B. conceived of the analysis, collected data, and wrote the manuscript. N.M.C. conducted the analysis and wrote portions of the Methods and Results sections of the manuscript, B.A. contributed data and critically revised the manuscript, C.C. contributed data and critically revised the manuscript, I.V.H. critically revised the manuscript, R.C.H. contributed data and critically revised the manuscript, J.M. critically revised the manuscript, B.J.Y. contributed data and critically revised the manuscript, A.H.K.-S. coordinated data collection across sites, oversaw analysis, and drafted portions of the Methods section, D.D.S. critically revised the manuscript, and C.W. oversaw research conception and design and critically revised the manuscript. G.T.L. contributed data, contributed to the conception of the analyses, and critically revised the manuscript.
References
- National Committee for Quality Assurance. HEDIS 2014 Measures. 2014; http://www.ncqa.org/hedis-quality-measurement/hedis-measures/hedis-archives. Accessed January 12, 2018.
- Harris AH, Humphreys K, Finney JW. Veterans Affairs facility performance on Washington Circle indicators and casemix-adjusted effectiveness. J Subst Abuse Treat. 2007;33:333–339.
- Green CA, Polen MR, Dickinson DM, Lynch FL, Bennett MD. Gender differences in predictors of initiation, retention, and completion in an HMO-based substance abuse treatment program. J Subst Abuse Treat. 2002;23:285–295.
- Acevedo A, Garnick DW, Lee MT, et al. Racial/ethnic differences in substance abuse treatment initiation and engagement. J Ethn Subst Abuse. 2012;11:1–21.
- Lee MT, Garnick DW, O'Brien PL, et al. Adolescent treatment initiation and engagement in an evidence-based practice initiative. J Subst Abuse Treat. 2012;42:346–355.
- Bensley KM, Harris AHS, Gupta S, et al. Racial/ethnic differences in initiation of and engagement with addictions treatment among patients with alcohol use disorders in the veterans health administration. J Subst Abuse Treat. 2017;73:27–34.
- Weisner C, Mertens J, Tam T, Moore C. Factors affecting the initiation of substance abuse treatment in managed care. Addiction. 2001;96:705–716.
- Brown CH, Bennett ME, Li L, Bellack AS. Predictors of initiation and engagement in substance abuse treatment among individuals with co-occurring serious mental illness and substance use disorders. Addict Behav. 2011;36:439–447.
- Harris AH, Bowe T, Finney JW, Humphreys K. HEDIS initiation and engagement quality measures of substance use disorder care: impact of setting and health care specialty. Popul Health Manag. 2009;12:191–196.
- Mertens JR, Flisher AJ, Satre DD, Weisner C. The role of medical conditions and primary care services in five-year substance use outcomes among chemical dependency treatment patients. Drug Alcohol Depend. 2008;98(1–2):45–53.
- Burman ML, Kivlahan D, Buchbinder M, et al. Alcohol-related advice for Veterans Affairs primary care patients: Who gets it? Who gives it? J Stud Alcohol. 2004;65(5):621–630.
- Dom G, Wojnar M, Crunelle CL, et al. Assessing and treating alcohol relapse risk in liver transplantation candidates. Alcohol Alcohol. 2015;50(2):164–172.
- Satre DD, Knight BG, Dickson-Fuhrmann E, Jarvik LF. Substance abuse treatment initiation among older adults in the GET SMART program: effects of depression and cognitive status. Aging Ment Health. 2004;8(4):346–354.
- National Committee for Quality Assurance. Summary table of measures, product lines and change. HEDIS 2015, Volume 2 (p. 8) 2015; http://www.ncqa.org/Portals/0/HEDISQM/Hedis2015/List_of_HEDIS_2015_Measures.pdf. Accessed January 12, 2018.
- Agency for Healthcare Research and Quality (AHRQ). Engagement of alcohol and other drug (AOD) treatment: percentage of members who initiated treatment and who had two or more additional services with a diagnosis of AOD within 30 days of the initiation visit. 2015; https://www.qualitymeasures.ahrq.gov/summaries/summary/49778? Accessed January 17, 2018.
- Health Care Systems Research Network (HCSRN). VDW Data Model 2017. 2017; http://www.hcsrn.org/en/Tools%20&%20Materials/VDW/. Accessed January 12, 2018.
- National Cancer Institute. Cancer Research Network (CRN). 2017; https://www.crn.cancer.gov/. Accessed January 12, 2018.
- Wagner EH, Greene SM, Hart G, et al. Building a research consortium of large health systems: the Cancer Research Network. J Natl Cancer Inst Monogr. 2005;2005(35):3–11.
- Hornbrook MC, Hart G, Ellis JL, et al. Building a virtual cancer research organization. J Natl Cancer Inst Monogr. 2005;2005:12–25.
- Ross TR, Ng D, Brown JS, et al. The HMO Research Network Virtual Data Warehouse: A public data model to support collaboration. Egems. 2014;2:1049.
- Charlson ME, Charlson RE, Peterson JC, et al. The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients. J Clin Epidemiol. 2008;61:1234–1240.
- Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619.
- Healthcare Cost and Utilization Project (HCUP). Clinical Classifications Software (CCS) for ICD-9-CM. 2017; https://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp#examples. Accessed June 27, 2018.
- Weisner C, Mertens J, Parthasarathy S, Moore C, Lu Y. Integrating primary medical care with addiction treatment: A randomized controlled trial. JAMA. 2001;286(14):1715–1723.
- Mello MJ, Baird J, Lee C, Strezsak V, French MT, Longabaugh R. A randomized controlled trial of a telephone intervention for alcohol misuse with injured emergency department patients. Ann Emerg Med. 2016;67:263–275.
- Laine C, Hauck WW, Gourevitch MN, et al. Regular outpatient medical and drug abuse care and subsequent hospitalization of persons who use illicit drugs. JAMA. 2001;285(18):2355–2362.
- Rollnick S, Miller WR, Butler CC. Motivational Interviewing in Health Care: helping Patients Change Behavior. New York, NY: Guilford Publications, Inc.; 2008.
- Nead K. Take the HIV challenge. 2012; https://centerfortotalhealth.org/announced-today-the-hiv-challenge/. Accessed March 16, 2018.
- Bobb JF, Lee AK, Lapham GT, et al. Evaluation of a pilot implementation to integrate alcohol-related care within primary care. Int J Environ Res Public Health. 2017;14(9):pii: E1030.
- Umbricht-Schneiter A, Ginn DH, Pabst KM, Bigelow GE. Providing medical care to methadone clinic patients: referral vs on-site care. Am J Public Health. 1994;84:207–210.
- Gourevitch MN, Chatterji P, Deb N, Schoenbaum EE, Turner BJ. On-site medical care in methadone maintenance: associations with health care use and expenditures. J Subst Abuse Treat. 2007;32:143–151.
- Laine C, Lin YT, Hauck WW, Turner BJ. Availability of medical care services in drug treatment clinics associated with lower repeated emergency department use. Med Care. 2005;43:985–995.
- Ng Fat L, Cable N, Shelton N. Worsening of health and a cessation or reduction in alcohol consumption to special occasion drinking across three decades of the life course. Alcohol Clin Exp Res. 2015;39:166–174.
- Satre DD, Gordon NP, Weisner C. Alcohol consumption, medical conditions, and health behavior in older adults. Am J Health Behav. 2007;31(3):238–248.
- Shaper AG, Wannamethee G, Walker M. Alcohol and coronary heart disease: a perspective from the British Regional Heart Study. Int J Epidemiol. 1994;23:482–494.
- Metz TD, Rovner P, Hoffman MC, et al. Maternal deaths from suicide and Overdose in Colorado, 2004-2012. Obstet Gynecol. 2016;128:1233–1240.
- HEDIS. Summary Table of Measures, Product Lines and Changes. 2018; https://www.ncqa.org/Portals/0/HEDISQM/HEDIS2018/HEDIS%202018%20Measures.pdf?ver=2017-06-28-134644-370. Accessed April 14, 2018.