New Generation Atypical Anti-Psychotic Medications: Access to Care and Criminal Justice Outcomes John A. Pandiani, Ph.D. Sheila M. Pomeroy, M.S.
[email protected] [email protected] Chief of Research and Statistics Senior Research and Statistics Analyst Vermont Department of Developmental and Mental Health Services 103 South Main Street, Waterbury, Vermont 05671-1610 Steven M. Banks, Ph.D.
[email protected] Chief Statistician, The Bristol Observatory 521 Hewitt Road, Bristol, Vermont 05443
ABSTRACT
Fewer adults served by community mental health programs for people with severe and persistent mental illness got into trouble with the law after beginning atypical anti-psychotic medication than before. This impact was not uniformly felt, however. Men experienced decreased criminal justice involvement (from 15% during the six months before medication to 11% during the first six months receiving medication), while criminal justice involvement for women increased (from 3.0% to 3.6%) during the same time period. Access to atypical medication was substantially greater for people who had previously been in trouble with the law and the difference was greater among younger clients. This statewide study relied entirely on existing databases in conjunction with emerging statistical techniques that measure cross service sector involvement without reference to personally identifiable information.
Presented at the Twelfth Annual NASMHPD Conference on State Mental Health Agency Services Research, Program Evaluation, and Policy "Moving Toward Evidence-Based Systems of Care" February 11, 2002 - Renaissance Harborplace Hotel - Baltimore, Maryland This project was supported, in part, by a Center for Mental Health Services Sixteen State Indicator Pilot Project Grant #5HR1SM52625-03. This report reflects the views and opinions of the authors. It does not necessarily reflect the official position or policy of CMHS, SAMHSA, or HHS.
Utilization of new "atypical" anti-psychotic medications has been identified by the National Association of State Mental Health Program Directors (NASMHPD) as one of the, "evidence-based practices" that should be implemented by publicly funded community mental health programs. NASMHPD's framework includes rates of access to atypical anti-psychotic mediation for clients served by state mental health programs as an important indicator of the performance of systems of care because, "Organizations that are focused on providing quality services work to insure that consumers are receiving treatments that are consistent with defined 'best practices'. At least two sets of guidelines have been developed for the treatment of psychosis. Both consider new generation atypical antipsychotics to be preferable over older agents. New generation agents have demonstrated advantages in efficacy and, with the exception of clozapine, safety over older agents. Use of the agents may be an indicator of the degree to which consumers of the organization are receiving treatments that conform to best practices."1 Much of the research endorsing this practice pattern, however, is hospital based and the generalizability of the findings to community-based programs that serve a diverse clientele in a variety of settings is not clear. This presentation reports the findings of a statewide study of the integration of new generation atypical anti-psychotic medications into routine practice patterns in Vermont's statewide communitybased system of care for adults with severe and persistent mental illness. Two basic questions were addressed in this research. The issue of treatment outcomes was examined by comparing rates of criminal justice involvement before receipt of atypical medication to rates of criminal justice involvement during receipt of atypical medication for people "going on" atypical medication, and by comparing rates of criminal justice involvement for people who received atypical medications and those who did not. In addition, the issue of access to care was examined by comparing rates of utilization of atypical antipsychotic medication for people in different demographic categories. Finally, the rate at which people begin using atypical anti-psychotics after criminal justice involvement will be compared to the rate at which people without recent criminal justice involvement begin using atypical anti-psychotic medication. Subjects The subjects of this research include all 4,200 adults aged 18 to 60 who received Medicaid reimbursed services from one of Vermont's community mental health programs for adults with severe and persistent mental illness during January 1996 through June 2001. The study population included more women than men (54% vs. 46%). Almost half (47%) of the subjects were 41 to 60 years of age, one third (33%) were 30 to 40, and about one in five (21%) were in the 18 to 29 year age group. Recipients of Medicaid reimbursed services represented 90% of all individuals aged 18 to 60 who were served by these community programs during the study period. Forty-two percent of the individuals served by these programs have a primary schizophrenia diagnosis (schizophrenia, schizophrenoform, or schizoaffective), and 32% have a primary affective diagnosis (bipolar, major depression, or dysthymia). Fifty-five percent of the individuals in the subject population had received atypical medication during July 1997 through December 2000, the period covered by this study. Method This study involved a combination of traditional and new research methodologies. The research design conformed to a classic pre/post quasi-experimental design 2 and used unobtrusive measures to control both the cost and the reactive effects of measurement 3. Traditional record linkage technology and the recently developed technology of Probabilistic Population Estimation4 (PPE) were used in the data analysis. PPE is a sophisticated computer intensive statistical procedure for determining the number of individuals shared by data sets that do not include unique person identifiers. This project relied entirely on extracts from two statewide operational/administrative databases. Information on utilization of community mental health programs and atypical medications was obtained from Vermont's Medicaid paid claims database maintained by the Office of Vermont Health Access. Information on individuals charged with a crime was obtained from the Vermont Center for Justice Research database of all Superior Court cases. The Medicaid database provided two sets of targeted extracts. The first set of extracts included all individuals served by community programs for adults with severe and persistent mental illness during
the specified time period. The second set of extracts included all paid claims for atypical anti-psychotic medication. The community program data set was aggregated to the person level. The medication data set was aggregated to the person and calendar quarter level, and the two files were linked on the basis of Medicaid's unique person identifier. Each of the resulting records included the date of birth and gender of the community service recipient and the recipient's atypical medication status (yes/no) for each calendar quarter during the study period. The Vermont District Court database provided records of all criminal charges (including felonies and misdemeanors, but not traffic offences) filed during each calendar quarter. Each record in each of these files included the offence charged, the date of the charge, and the date of birth and gender of the person charged. For purposes of analysis, each individual in the subject group was placed into one of three categories with regard to their receipt of atypical medication during each fiscal and each calendar year in the study period. Individuals were categorized as "going on" atypical medication if they did not receive atypical medication during the first two quarters but did receive atypical medication during each of the following two quarters. These were the primary subjects of this study. For purposes of comparison, two additional groups were identified. Individuals were categorized as "staying on" atypical medication if they received medication during all of the four quarters. Individuals were categorized as "staying off" atypical medication if they did not receive medication during any of the four quarters. In order to determine the degree to which "going on" atypical anti-psychotic medication was related to the rate at which mental health service recipients got into trouble with the law, the rate at which service recipients were charged with a crime during two time periods was determined. The first time period was the first half of the year, when they did not receive medication. The second time period was the second half of the year, when they did receive medication. For purposes of comparison, rates of criminal justice involvement during the first half and the second half of each year were determined for individuals "staying on" and for individuals "staying off" atypical medication. The resulting rates of criminal justice involvement were averaged for the six study years examined in this study. For individuals "going on" atypical medication, the impact of the medication was measured by comparing rates of criminal justice involvement during the first half of each study year (the quarters with no atypical medication) to the second half of the study year (the quarters with atypical anti-psychotic medication). Similar analyses were also conducted for each of the two comparison groups. Rates of criminal justice involvement are based on the overlap between the mental health and the criminal justice data sets. Because the Medicaid and the criminal justice data sets do not share unique person identifiers, Probabilistic Population Estimation was used to determine the rate of criminal justice involvement for each group of mental health service recipients. Probabilistic Population Estimation is a statistical procedure that provides valid and reliable measures of the size and overlap of data sets that do not include unique person identifiers. These estimates are based on a comparison of the distribution of dates of birth in the data sets to the known distribution of dates of birth in the general population. A technical description of the procedure and a list of published papers that describe and demonstrate the use of the procedure are provided at the end of this handout. Findings Results of this analysis indicate that being charged with a crime, our criminal justice treatment outcome, is associated with utilization of atypical anti-psychotic medication, but that the magnitude and direction of this association varies with age and gender. Access to atypical anti-psychotic medication varies with age (but not gender) and is related to previous criminal justice involvement. Access to Care Access to atypical medication was not impacted by gender, varied slightly with age, and was substantially influenced by prior criminal justice involvement. Overall, more than half of all adults in the subject population received atypical medication during the study period. Men and women received atypical medication at about the same rate (54% and 56% respectively). Young adults in the 18 to 29 year age group were somewhat less likely to receive atypical medication than adults in the 30 to 40 and 41 to 60 age groups (49% vs. 58% and 57% respectively).
Criminal justice involvement had a much greater impact on access to atypical medication than client demographics. Compared to others, the community mental health clients who were charged with a crime were 2.6 times as likely to "go on" atypical medication during the following quarter. This increased access to atypical medication for people in trouble with the law was evident for both men and women and for every age group. The difference between people in trouble with the law and others was greatest for young adults (4.4 times as likely) and least for older adults (1.5 times as likely); men and women in trouble with the law experienced increased access to atypical medication at similar rates. Treatment Outcomes Rates of criminal justice involvement for individuals "going on" atypical anti-psychotic medication decreased after going on atypical medication (from 7.9% to 6.6%) although this was not true for clients in all age/gender categories. People who did not receive atypical anti-psychotic medications were charged with a crime at about the same rate as people who received atypical anti-psychotics during the entire study period, but there were significant differences between the two groups in some age/gender groups. Men experienced a significant and substantial decrease in criminal justice involvement (from 15% to 11%) after going on atypical medication. This decrease, however, was only evident in the 30 to 40 year age group, where criminal justice involvement decreased by almost 50%, from 20.8% during the six months before going on atypical medication to 11.2% during the first six months of receiving atypical medication. There were no statistically significant differences for men in our other two age groups. Interestingly, men in the 30 to 40 year age group who consistently received atypical medication had significantly lower criminal justice involvement than men who received no atypical medication (3.9% vs. 6.6%), but men in the 18 to 29 year age group who consistently received atypical medication had significantly higher criminal justice involvement than men who received no atypical medication (9.1% vs. 5.7%). There were no differences in criminal justice involvement between men in our oldest age group who consistently received medication and those who did not receive atypical medication at all. Consumers Charged with a Crime and Use of Atypical Anti-Psychotic Medication 16% Time 1
Time 2
12%
8%
4%
0% No
No/Yes Yes No No/Yes Yes Total Male No = No atypical anti-psychotic medication No/Yes = Received atypical anti-psychotic medication during time 2 only Yes = Received atypical anti-psychotic medication during both time periods
No
No/Yes Female
Yes
Women, on the other hand, experienced a significant (but less substantial) increase in criminal justice involvement (from 3.0% to 3.6%) after going on atypical medication. This increase was most evident in the 18 to 29 year age group, where criminal justice involvement increased by more than 35%, from 6.4% during the six months before going on atypical medication to 8.7% during the first six months
of receiving atypical medication. There was also a statistically significant increase for women in the 30 to 40 age group. The rate of criminal justice involvement for women in the 41 to 60 age group decreased at a statistically significant rate from 1.2% to 1.0%. Interestingly, women in both the 18 to 29 year and the 30 to 40 year age group who consistently received atypical medication had significantly higher criminal justice involvement than women who received no atypical medication (5.9% vs. 2.9% and 2.8% vs. 2.1%, respectively). There were no differences in criminal justice involvement between women in our oldest age group who consistently received medication and those who did not receive atypical medication at all. Discussion This project has demonstrated the utility of existing data sets, in conjunction with sophisticated statistical tools, for monitoring the implementation and evaluating the outcomes of evidence-based practice in community settings. The results of this project clearly identified populations that have benefited from atypical medications and populations that have not. Men in the 30 to 40 age group benefited most in terms of reduced criminal justice involvement. Women in the 18 to 29 and 30 to 40 age groups, by contrast, actually experienced increased criminal justice involvement that was associated with atypical anti-psychotic medication. We believe the methodology that was used in this study provides a model for efficient and effective evaluation of the implementation of evidence-based practice in community settings. Because this approach uses existing data resources, it is much more economical than original data collection and it can be used to continually evaluate changing practice patterns as they occur. Different treatment models can be compared in terms of both access to care and treatment outcomes. This methodology facilitates the measurement of the long-term functional outcomes that are most relevant to community settings while preserving anonymity and privacy. This is particularly important in the context of increasing regulation and limitation of access to personally identifiable information because the technology is designed to use anonymous computer records 5. Next Questions These findings suggest a number of important areas for further inquiry regarding the efficacy of psychotropic medication in community settings. The following questions suggest directions for further research that we believe would be particularly valuable. 1) Do similar patterns exist in other geographical locations? Similar research should be conducted in other states (and nations), and include urban as well as rural areas. Future research should also include more ethnically diverse areas and should test the impact of race/ethnic variation on the results reported here. 2) Do similar patterns exist for other outcomes? Future research should include other outcome measures such as rates of hospitalization for behavioral health care, other measures of criminal justice involvement (e.g. arrest, and incarceration), as well as positive outcomes such as employment rates. 3) Are the observed patterns consistent across all atypical medications (and combinations of medication) or are they specific to particular medications (or combinations of medications)? Future research should include a narrowing of focus to include more specificity regarding the medications provided to recipients of community-based services. References 1
National Association of State Mental Health Programs Directors. (1998) Performance Measures of Mental Health Systems. Arlington, VA. National Association of State Mental Health Program Directors. http://www.rdmc.org/nri/firstpage.htm 2 Campbell DJ & Stanley, JC (1963) Experimental and Quasi-Experimental Designs for Research. Chicago, Rand McNally. 3 Webb EJ, Campbell DT, Schwartz RD & Sechrest, L (1966) Unobtrusive Measures: Nonreactive Research in the Social Sciences. Chicago: Rand McNally. 4 Banks SM & Pandiani, JA (2001) Probabilistic population estimation of the size and overlap of data sets based on date of birth. Statistics in Medicine. 20:1421-1430. 5 Pandiani JA, Banks SM & Schacht LM (1998) Personal privacy vs. public accountability: A technological solution to an ethical dilemma. Journal of Behavioral Health Services and Research. 24 (4): 33-44.
METHODOLOGICAL NOTE PROBABILISTIC POPULATION ESTIMATION Probabilistic Population Estimation is a statistical procedure that determines the number of people (with known confidence intervals) who are represented in data sets that do not contain unique person identifiers. Probabilistic Population Estimation uses information on the distribution of birth dates in a data set to determine the number of people represented in the data set. The number of people necessary to produce the number of birthdays observed in a single birth year cohort, for instance, would be calculated using the following formula:
Pj (l j ) = ∑ l
i =1
365 365− i
where “Pj” is the number of people and ”i” is the number of birth dates observed. Similar logic is used to determine the number of people who appear in more than one data set. The table below provides illustrative results of Probabilistic Population Estimation for populations of specified size.
Population Estimates for Specified Numbers of Birth Dates Within a Year Birth Dates 1 10 20 50 100
Number of People 1.003 ± .103 10.15 ± .776 20.6 ± 1.54 54. ±4 117. ±9
Birth Dates 180 250 300 330 360
Number of People 249 ± 20 423 ± 38 632 ± 64 860 ± 101 1630 ± 325
POPULATION OVERLAP In order to probabilistically determine the number of people shared across data sets that do not include a common person identifier, the sizes of three populations are determined and the results are compared. The number of people in each of the original data sets are the first two populations. The number of people in a data set that is formed by combining the two original data sets is the third data set. The number of people who are shared by the two data sets is the difference between the sum of the numbers of people represented in the two original data sets and the number of people represented in the combined data set. This occurs because the sum of the number of people represented in the two original data sets includes a double count of every person who is represented in both data sets. The number of people represented in the combined data set does not include this duplication. The difference between these two numbers is the size of the duplication between the two original data sets, the size of the caseload overlap. In terms of mathematical set theory, the intersection of two sets is the difference between the sum of the sizes of the two sets (A+B) and the union of the two sets (A∪B): (A ∩ B) = (A + B) - (A ∪ B).
Related Reading
A Global Measure of Access to Mental Health Services for a Managed Care Environment. Journal of Mental Health Administration, Summer 1997. (Pandiani, Banks, & Gauvin) A Methodology for Probabilistically Estimating Caseload Size and Overlap. The Evaluation Center @HSRI, January 1999 (Banks & Pandiani) A Risk Adjusted Measure of Hospitalization Rates for Evaluating Community Mental Health Program Performance. Administration and Policy in Mental Health, March, 1999. (Pandiani, Banks, Schacht, & Gauvin) After Children’s Services: A Longitudinal Study of Significant Life Events. Journal of Emotional and Behavioral Disorders, 2001 Vol. 9 #2 (Banks, Pandiani, & Schacht) Age and Mortality Among Problem Drinkers. Addiction. August 2000 (Banks, Pandiani, Schacht, & Gauvin) Approaches to Risk Adjusting Outcome Measures Applied to Criminal Justice Involvement after Community Service. Journal of Behavioral Health Services and Research. 2001 Vol. 28 #3 (Banks, Pandiani, & Bramley) Bed Closures and Incarceration Among Users of VA Behavioral Health Services in Upstate New York. Mental Health Services and Research, October 2000 (Rosenheck, Banks, Pandiani, & Hoff) Caseload Segregation/Integration: A Measure of Shared Responsibility for Children and Adolescents. Journal of Emotional and Behavioral Disorders, Summer 1999 (Banks, Pandiani, & Schacht) Consumer Satisfaction and Treatment Outcomes: Dissatisfaction with Mental Health Services and Incarceration after Treatment. Administration and Policies in Mental Health, 2001, Vol. 29, #2. (Pandiani, Schacht, & Banks) Does Closing Inpatient Beds in One Public Mental Health System Result in Increased Use of Inpatient Services in Other Systems? Psychiatric Services, 2000 Vol. 2 #4. (Rosenheck, Banks, & Pandiani) Elevated Risk of Being Charged with a Crime for People with a Severe and Persistent Mental Illness. Justice Research and Policy, Fall 2000 (Pandiani, Banks, Clements, & Schacht) Personal Privacy vs. Public Accountability: A Technological Solution to an Ethical Dilemma. Journal of Behavioral Health Services and Research, November 1998. (Pandiani, Banks, & Schacht) Practice Patterns and Hospitalization Rates: A Statewide Program Evaluation. Administration and Policy in Mental Health, September, 1998. (Banks, Pandiani, Gauvin, Reardon, Schacht, & Zovistoski) Probabilistic Population Estimation of the Size and Overlap of Data Sets Based on Date of Birth. Statistics in Medicine, 2000. (Banks & Pandiani). The Use of State and General Hospitals for Inpatient Psychiatric Care. American Journal of Public Health, March 1998. (Banks & Pandiani) Utilization of Local Jails and General Hospitals by State Psychiatric Center Patients. The Journal of Behavioral Health Services and Research, November 2000 (Banks, Stone, Pandiani, Cox, & Morchauser) Using Existing Data Bases to Measure Treatment Outcomes. In Developing Outcome Strategies in Children's Mental Health. Edited by Hernandez, M and Hodges, S. Paul H. Brookes Publishing Co, Baltimore 2001 (Banks & Pandiani) Using Incarceration Rates to Measure Mental Health Program Performance. Journal of Behavioral Health Services and Research, August 1999. (Pandiani, Banks, & Schacht)