Volume 1, Issue 2

Fall 1999    

At an accelerating rate we are hearing more and more about "best practices." While others act as though they understand the concept, we wonder if we are the only ones who are not certain what the term actually means. Is it just a new catch phrase? Does it imply that there is only one "best" approach? (After all, the phrase is best practice not good practice, not one of a number of good practices, and not standard of care.) Since we have included Best Practices in the title of our publication, we decided we had better take a stab at addressing what we think it means.

First, we doubt that there are best practices out there waiting to be discovered. An approach to practice is not like a pot of gold waiting to be found. There are approaches to providing treatment and service delivery that are better than other approaches. Second, there are approaches that are likely better for some subpopulations and not as effective for others. Mental health treatment progresses through a series of successive approximations. We do not go directly from bad practice to a best practice. The General Services Administration (GSA) defines best practices as "good practices that have worked well elsewhere." A national publication, Outcomes Accountability Alert, recently published an article entitled, "How to incorporate evidence-based methods into routine clinical care." We agree with GSA, and are particularly fond of the concept of evidence-based methods as opposed to best practice. (Too bad we have already spent so much time thinking about the title for this publication!) The one idea that we would like to add, however, is population specificity. We would like to propose that a best or evidence-based practice is a specific, replicable method that leads to successful outcomes for specific populations. All too often in the mental health field innovative approaches that are successful for one population are applied to populations for whom the method was not designed. And like trying to fit a square peg into a round hole, the applications generally do not fit - no matter how hard we hammer away. We should also be humble enough to recognize that best practices exist in a time continuum. Today's best practice is likely to be not much more than historic perspective for tomorrow's best practice. Surgical practices; medications for a variety of physical and psychiatric illnesses; and mental health treatment in general are better today than they were yesterday. It is likely, however, that these same practices will be even better tomorrow than they are today. The best practice is the best that we can do (in terms of technology and implementation) at a point in time. It is not a final product.

In 1994, Dr. William Glazer authored an article published in Psychiatric Services entitled "What are 'Best Practices?': Understanding the Concept." In the article he indicated that we needed to be aware of laboratory findings (efficacy), and apply those findings in the real world (effectiveness). In a 1998 follow up article, Dr. Glazer expressed concern that not much progress was taking place in best practice investigations. He cited obstacles, including the fear of taking risks in developing and investigating innovative approaches to treatment, as well as the lack of resources and research/evaluation expertise in the community. We hope that a forum like Outcomes, Innovations & Best Practices might serve, in some small way, to add to the available resources. We hope also that Outcomes, Innovations & Best Practices provides a forum for providers to discuss and share approaches to treatment, rehabilitation, prevention and (very importantly) methods of evaluating effectiveness.

To share your thoughts on the issue, contact Dr. Gentile at edgen@cpsa-rbha.org v

Most of us are regular recipients of information telling us What to do and How to do it. But since the best approach to learning is to learn from failure, presented below is a tongue-in-cheek list of 11 strategies guaranteed to lead to failure in the world of conducting outcome evaluations. Be careful, because if you avoid these eleven steps you might be well on your way to completing a successful outcome study.

• Do not consult with your stakeholders to find out what information they think is useful. Since you are the expert, the more you force your ideas on others, the more they will respect the data you have. If you feel you must consult with stakeholders, consult only with those who you believe are "above you" in some administrative hierarchy. Those below you, the line staff who will be actually gathering the data, will be forced to do what you want, so you don't need to solicit their ideas of what is helpful for improving treatment.

• Consumers are a special class of stakeholders with whom you should not consult. This includes not only the individual receiving services, but also family members. If they don't like it, they can go elsewhere.

• Choose outcome tools that are complex. The tools should be difficult to administer, score, and interpret. Selecting expensive tools is a bonus. Outcome tools that need computer scoring or advanced training to use (or both) should be high on your priority list.

• Do not train staff on any aspect of your outcome studies project. If they think they understand the project, they might offer suggestions for improvement.

• Outcome instruments need not have clinical utility. Providing direct care staff with an instrument that might be useful will only confuse them.

• Clients and family members do not need to know the results of the survey forms they complete. Nor do they need to be bothered with aggregate results.

• Make sure outcome tools provide limited information and that it is useful to only a small group. Evaluation questions should have little relevance to the services and treatment that are actually provided. Otherwise, there is a temptation to use the results to improve the system of care. Whatever you do, don't use the results to make changes.

• Report back to the stakeholders as infrequently as possible. What they don't know can't hurt them.

• If you must report, make sure that your reports are long and obtuse. Give detailed information about statistics that most of your readers don't care about or understand. Bonus points are given for verbose discussions of one-tail vs. two-tailed tests, methods used to calculate degrees of freedom etc. Use "etc." a lot-particularly if the readers do not know what other examples would be. Use abbreviations and acronyms. Fill your reports with references to academic theories and studies.

• Avoid humor. This is serious science. Use of humor might help readers and listeners appreciate and understand what it is that you are saying.

• An atmosphere of paranoia will help motivate the staff. Let rumors spread that the results of your outcome studies might be used as against them.

Over the past few years, there has been a great deal of discussion concerning the new antipsychotic medications. In order to help ensure that the discussion in Southern Arizona remains grounded on published research, this article has been prepared to provide a bibliography of some of the more important findings. This review is not intended to be comprehensive. It is written for those who might not be current with the literature. Additionally, while there are possible differences among the newer medications, exploration of that literature is beyond the scope of this overview. Rather this overview focuses on the issue of the new medications as a class vs. the older medications.

The literature that addresses the new medications can be classified as falling into three basic areas:

• Pharmaceutical properties that make the new medications atypical
• Effectiveness of the new medications
• Cost and cost offset of the new medications

The term "atypical" refers to the characteristics of the medications that significantly reduce the incidence of side effects seen with traditional antipsychotic agents such as haloperidol. In essence, the atypical medications have a low affinity for neuronal receptors responsible for motor effects and a high affinity for receptors associated with psychiatric symptoms. The term "atypical" was originally associated with clozapine, but is also associated with olanzapine, risperidone, and most recently quetiapine. At higher doses, however, these newer atypicals can result in side effects. An excellent overview article addressing pharmaceutical properties and other issues, including differences among the atypicals, can be found in:

Jibson MD, Tandon R: New atypical antipsychotic medications. Journal of Psychiatric Research 1998; 32: 215-218.

While there are a number of specific outcome measures that have been used in research settings, they fall within four basic categories: Positive Symptoms, Negative Symptoms, Side Effects and Quality of Life.

Positive symptoms: In studies addressing positive symptoms, the atypicals have been demonstrated to be superior to haloperidol in: a) decreasing psychotic symptoms (as measured by instruments such as the "Brief Psychiatric Rating Scale" and the "Positive and Negative Syndrome Scale of Schizophrenia"), b) reducing hospitalization rates, c) reducing rates of relapse, and e) improving cognitive processing. Studies addressing positive symptoms include:

Luchins DJ, Hanrahan P, et al.: Initiating clozapine treatment in the outpatient clinic: service utilization and cost trends. Psychiatric Services. 1998; 49: 1034-1038.

Tollefson GD, Beasley CM, et al.: Olanzapine versus haloperidol in the treatment of schizophrenia and schizoaffective and schizophreniform disorders: Results of an international collaborative trial. American Journal of Psychiatry 1997; 154: 457-465.

Marder S, Meiback R: Risperidone in the treatment of schizophrenia. American Journal of Psychiatry 1994; 151: 825-835.

Green MF, Marshall BD, et al.: Does risperidone improve verbal working memory in treatment-resistant schizophrenia? American Journal of Psychiatry 1997; 154: 799-804.

Negative symptoms: Negative symptoms are somewhat less obvious, but nonetheless extremely important, and include lethargy, disinterest in life and a general atrophy of life skills. Five of the better studies that report significant effects as measured by the "Scale for the Assessment of Negative Symptoms" are:

Marder S, Meiback R: 1994 (see article cited above).

Tollefson GD, Sanger TM: Negative Symptoms: A Path Analytic Approach to a Double Blind, Placebo- and haloperidol-controlled clinical trial with olanzapine. American Journal of Psychiatry 1997; 154: 466-474.

Glazer WM: Clinical outcomes of pharmacotherapy for schizophrenia and implications for health economics. Journal of Clinical Psychiatry 1997; 15: 22-23.

Meltzer HV: Outcome in Schizophrenia: beyond symptom reduction. Journal of Clinical Psychiatry 1999; 60: 3-8.

Lane HY, Liu CC et al.: Risperidone for exclusively negative symptoms. American Journal of Psychiatry 1999; 156: 335-340.

Side effects: The primary side effects of the older medications are extrapyramidal symptoms (involuntary muscle movements), akathisia (restlessness), anxiety, dystonia (acute muscle spasms) and parkinsonism. The most significant potential side effect of traditional medications are the life-long irreversible consequences of tardive dyskinesia and tardive dystonia. Data demonstrates quite clearly that the atypical medications result in significantly reduced side effects. This is extremely important when it is recognized that one of the major reasons for medication non-compliance is discomfort with side effects. Studies discussing side effects include:

Weiden P, Aquilar R, et al.: Atypical antipsychotic drugs and long-term outcome in schizophrenia. Journal of Clinical Psychiatry 1996; 57: (Supplement).

Barnes TR, McPhillips MA: Novel antipsychotics, extrapyramidal side effects and tardive dyskinesia. International Clinical Psychopharmacology 1998; 13: 13-57.

Fernandez HH, Friedman JH: The role of atypical antipsychotics in the treatment of movement disorders. CNS Drugs 1999; 11: 467-483.

Chouinard G: Effects of risperidone in tardive dyskensia: an analysis of the Canadian multicenter risperidone study. Journal of clinical psychopharmacology. 1995; 15: 36-44.

Casey DE: The relationship of pharmacology to side effects. Journal of Clinical Psychiatry monograph 1997; 58: 55-62.

Quality of life: Quality of life is often referred to as a "softer measure." It is a subjective, difficult-to-measure concept and addresses the extent to which the individual does or does not suffer as a consequence of his or her illness. The less the suffering (based on issues such as relationships with friends with family, medication side effects, participating in normal activities, etc.) the better the quality of life. Studies addressing quality of life include:

Franz M, Lis S, et al.: Conventional versus atypical neuroleptics: subjective quality of life in schizophrenic patients. British Journal of Psychiatry 1997; 70: 422-425.

Chouinard G, Albright PS: Economic and health state utility determinations for schizophrenic patients treated with risperidone or haloperidol. Journal of Clinical Psychopharmacology 1997; 17: 298-307.

Glazer, WM: Atypical Antipsychotics and Formulary Decisions. The American Journal of Managed Care. 1998; 4: 8-14.

Aronson SM: Cost Effectiveness and Quality of Life in Psychosis: The pharmacoeconomics of risperidone. Clinical Therapeutics 1997; 19: 139-147.

The two cost issues are a) the actual acquisition cost of the medications and b) consequent reduction in utilization of other services (hospitalization, crisis services, etc.). The atypical medications are anywhere from 15 to 35 times more expensive than haloperidol, depending upon the dosage and specific medication. A number of articles indicate quite clearly, however, that although the atypicals are more costly per unit dose, the cost differences are made up for by their effectiveness. While this finding has been replicated numerous times, we should be cautious in Arizona where there is already an extremely low rate of hospital utilization compared to other localities. With such a low base rate, the cost offset associated with the atypicals may be less or even nonexistent. Studies discussing cost and cost offset include:

Hargreaves WA, Shumway M: Pharmacoeconomics of antipsychotic drug therapy. Journal of Clinical Psychiatry 1996; 9: 66-75

Albright P, Livingston S, Keegan D et al.: Reduction of healthcare resource utilization & costs following the use of risperidone for patients with schizophrenia previously treated with standard antipsychotic therapy. Clinical Drug Investigations 1996; 11: 289-299.

Fitchner CG, Hanrahan P, et al.: Pharmacoeconomics studies of atypical antipsychotics: Review and perspective. Psychiatric Annals 1998; 28: 381-396.

Glazer WM, Johnstone BM: Pharmacoeconomic evaluation of antipsychotic therapy for schizophrenia. Journal of Clinical Psychiatry 1997; 58: 50-54.

While the preponderance of the data is positive, some research has certainly indicated that the new medications are not a panacea. In one study, only one-third of those started on risperidone was still taking the medication two years later. The reasons for discontinuation were similar to those that we have seen with the older medications: side effects and non-response to the medication. While the data was collected prior to more recent modifications in dosing strategies, the findings remind us that as advanced as the new medications are, we will still be progressing by successive approximations (see the lead article in this issue of Outcomes, Innovations & Best Practices). The data described above is from:

Binder RL, McNiel DE et al.: A naturalistic study of clinical use of risperidone. Psychiatric Services 1998; 49: 524-526.

For anyone interested in having primary source information, there are three excellent articles. The first is the Hargreaves article cited above in the Cost and cost offset section. A second article, by Dr. Jeffrey Lieberman addresses each of the medications. In the third article, Dr. Peter Buckey presents a current and complete overview. If you would like copies but don't have ready access to a library, please let us know and we will make sure that you get a reprint.

Lieberman JA: The decade of the brain. National Alliance for the mentally ill. 1997; 8: 1-15.

Buckey PF: New antipsychotic agents: emerging clinical profiles. Journal of Clinical Psychiatry 1999; 60: 12-17. v

If you are McDonalds, geoaccess revolves around the best locations for hamburger stands. The specific questions that a management team might ask include:

• Are there areas of the community that will support a new hamburger stand better than other parts of the community?
• How many miles between hamburger stands is too many miles?
• How many miles is too few miles?

Some of the geoaccess data that might be used to address such questions includes income level of various geographic areas of the community, the number of families that reside within the vicinity of a particular location and the number of automobiles that pass various locations during the day.

In public mental health, the parallel geoaccess questions revolve around the best locations for intake, treatment and rehabilitation sites. Specific questions include "Are there areas of the community that are in greater need of behavioral health services?", "Should certain types of services or programs be placed in specific geographic areas of the community?" "Are there areas of the community that are under served?" and "What type of public transportation is available to various sites?"

In this, the first of a two-part series to be published in Outcomes, Innovations & Best Practices, we will address:

• The types of data and information that should be considered in geoaccess analysis.

• Where one obtains geoaccess data.

In the next issue of Outcomes Innovations & Best Practices, we will discuss the process of presenting (ranging from colored pins on a wall map to sophisticated Geographic Information Systems or GIS software) and using geoaccess information.

Types of data and information that should be considered for geoaccess analysis

There are three types of geoaccess data/information that are important in the planning process:

• Demographic/social indicator data
• Utilization data
• Current facility information

Before discussing the variables that should be considered and where one might go to find data, the importance of using comparable geographic units (such as zip code, census tract and exact address) cannot be overstated. The reason for the compatibility is straightforward. If one were to imagine superimposing geoaccess maps on top of each other, you would certainly want comparable boundaries. Two maps, one portraying poverty by census tract and another portraying customer residence by zip code might be very difficult to use together. This is because customers that live in one zip code might live in four or five different census tracts, and each census tract might have a different poverty rate.

While exact address maps are compatible with both zip code and census tract maps, and are quite appropriate for mapping many variables (member utilization, achievement scores for specific schools, etc.), census tract (or even smaller census blocks) is the most appropriate unit of measure for mapping demographic and social indicator variables. The rationale for using census tract rather than zip code is that since the Bureau of Census (and other sources such as state and local health departments) collects and reports data by census tract, we have more knowledge about a specific area if we know which census tract it is in than if we know the zip code. While there are occasions when data are available only by zip code or other boundary unit (local law enforcement data are often reported according to "beat" or jurisdiction), the rule of thumb is to use census tract data when possible.

Important Demographic and Social Indicator Data: Some of the more important variables include Age (various breakdowns), Ethnicity, Gender, Population Density, Population Increase (Decrease), Income (various breakdowns), Number of Families, Rent as a Percent of Income, Criminal Justice Information (various breakdowns), Education (various breakdowns, including achievement scores and graduation rates), Divorce Rate, Unemployment Rate, Alcohol and Tobacco Use During Pregnancy, Teen Pregnancies, Juveniles in the Criminal Justice System, Number of Adults Unable to Speak English, Households Without an Automobile or Telephone Available, Mothers in the Labor Force, Single Head of Households, Households With Public Assistance Income, Crowded Housing, Causes of Deaths (various breakdowns) and Child Abuse.

The demographic and social indicator variables cited above can be obtained from the Bureau of Census, county health departments, county planning departments, various criminal justice entities (i.e., local police, sheriff, Juvenile Court, Department of Public Safety, and FBI), University Geography Departments, Department of Economic Security, Departments of Education and local school districts.

Other variables such as the number of liquor stores, new housing starts and number of churches can be obtained from a variety of sources including phone books, chambers of commerce and real estate databases.

It is important to note that in collecting and using social indicator data, much of the data should be analyzed in terms of both actual numbers and standardized rates. That is, while one might want to know which areas of the community have a greater number of alcohol related deaths, it is also important to know which areas of the community have the greater number of alcohol related deaths per 1000. For example, while census tract 22.02 might have twice as many alcohol-related deaths as census tract 44.03, tract 22.02 might also have five times the population base. Thus while the number of alcohol related deaths in tract 22.02 might be relatively high, the comparative rate might be low. Some social indicator data is available in terms of standardized or per capita numbers; for other sets the standardized rate must be calculated.

Internet Sites for Obtaining Data: Anyone who deals with internet sites is aware that once you jump in, the links between sites are many, and familiarity with only a few sites will allow you to tap into many others. Here are a few of our favorites:

• www.census.gov (certainly one of the more important sites)
• www.ciesin.org
• http://wonder.cdc.gov/
• www.ahcccs.state.az.us
• www.oseda.missouri.edu
• www.bea.doc.gov/
• www.icpsr.umich.edu
• www.whitehouse.gov/fsbr/demography.html
• http://fisher.lib.Virginia.EDU/ (From the home page, select libraries and electronic centers, then select geospatial and statistical data center. This is a great site with links to other data centers).

In addition to starting with specific web sites, using search engines such as Yahoo and Excite can yield additional valuable resources. Some of the more useful terms to use in searching the web include census maps, census data, crime data and health data.

As stated above, you will likely find that the links between sites are numerous and the amount of information plentiful.

Utilization Data: Utilization data is generally going to be agency-specific and there is nowhere to go for the data other than the agency's own data system. The types of data that are needed include number of individuals served and units of service (broken down by various distributions of age, ethnicity, presenting problem, and diagnosis etc). It is important with utilization data, as with social indicator data, to have both actual and standardized rates. And as indicated above, utilization data can be plotted either by address or census tract.

Current Facility Location: The most readily accessible information needed for geoaccess analysis is an inventory of current facilities. This information is obviously mapped in terms of exact address. In plotting facility location, one can also create maps identifying distances to facilities by using concentric circles.

Inserted in this issue of Outcomes, Innovations & Best Practices are two example maps that can be used in geoaccess analysis. Map 1 is a Metropolitan Pima County social indicator map (the percent of children living below the federal poverty level) with superimposed concentric circles identifying five-mile distances to CPSA Child/Adolescent Intake Sites. As you review the map, hopefully the wheels are turning and you are wondering about maps describing "areas of residence for enrolled children," or "the relationship between enrollment and poverty." Map 2 is also a social indicator map (the percent of households without an automobile"), but rather than just portraying Pima County, it demonstrates how geoaccess mapping can be used for planning on a state-wide basis.

We know that this is a lot to digest and would be happy to assist with questions that you might have. For more information, contact Dr. Berren at miber@cpsa-rbha.org or Melanie Brickman at melaniebrickman@yahoo.com

In the next issue of Outcomes, Innovations & Best Practices, we will discuss the specifics of turning raw data into geoaccess maps. v

CPSA publishes Outcomes, Innovations & Best Practices as a forum for providers of behavioral health services in Southern and Southeastern Arizona to share outcomes, innovative approaches to service delivery and other current research and needs assessment information with recipients of service, family members, stakeholders, and other providers. In order to facilitate the learning and improvement process, we encourage the sharing of approaches to service delivery that have been successful, as well as ones that have not been as successful. We firmly believe that best practices are developed through successive approximations.

Outcomes, Innovations & Best Practices is also being published as a vehicle for sharing the outcomes and innovative models that are being produced both elsewhere in Arizona and nationally.

If you received a "hand me down" copy of Outcomes, Innovations & Best Practices, and would like future copies mailed directly to you, please call Yolanda Hernandez at (520) 325-4268, or (800) 959-1063.

Also, if you did not receive a 3-ring binder for storing copies of Outcomes, Innovations & Best Practices, and would like one, please let us know and we will send you one.

In order to ensure that Outcomes, Innovations & Best Practices is a community based effort, we have put together an impressive Editorial Advisory Board. The advisory board will play an important role in reviewing submissions, and providing input concerning the focus of future issues.

Michael Berren, Ph.D. of CPSA

Linda Arzoumanian, Ed.D. of CODAC

Robyn F. Cruz, Ph.D. of COPE

Rick Duran, M.A. of CPSA

Cristóbal Eblen, Ph.D. of CPSA

Patricia E. Penn, Ph.D. of La Frontera Center, Inc.

Karl Sachs, Psy.D. of Arizona's Children Association

Michelle Stewart, M.A. of the University of Arizona

Yolanda Hernandez of CPSA