The evaluation of clinical data at most facilities is complicated by the fact that the treatment occurs in a real life setting without any experimental controls. Consequently there are many uncontrolled relevant variables that will significantly influence the outcome of your study. These variables often are not under the control of the clinician and may not even be related to the treatment milieu.
In the "real world" there are typically no control groups' baselines and post-treatment measures to compare against. Obviously this limits the reliability of the statistical analysis and the conclusions that can be drawn from it. This disadvantage can be offset to some degree by developing a large historical database of treatment and outcome data. The larger the sample size and the greater its consistency, the stronger the case that can be made for a causal relationship between treatment and effect.
One way to rapidly increase your treatment population sample is to pool data with similar types of providers. This may be done by developing a common database with other providers in your network, with other providers that receive a significant degree of revenue from the same funding source, or through commercial utilization review and outcome monitoring consulting firms. In most cases it is appropriate to develop a separate data repository or data warehouse for these data.
The assessments can be conducted verbally or using a written instrument, or entered directly into the program. The entire assessment instrument can be kept on-line or summary indices can be entered. The various outcome measures can then be correlated with interventions, clinicians, programs and provider's organizations to determine clinical effectiveness. Predictability can be improved through increasing sample size and by controlling for variables that influence the dependent measures.
By comparing treatment parameters with outcome measures a "dose-response" relationship may be established. "Dose-response" relationships may be created by comparing level of care (treatment intensity) and length of stay with outcome for a specific diagnostic category. The system can be used to help determine which treatment protocols are most beneficial and determine the optimal treatment time for specific presenting problems and diagnostic categories. With a fully integrated MIS, treatment cost can be determined and cost-benefit analyses conducted.
If multiple providers are going to participate in the design and support of a central database it is important that the process be standardized as much as possible to reduce variability. The same instruments should be administered in the same way, at the same time, etc. All definitions, treatment, data collection procedures, diagnostic and outcome categories and codes should be standardized across all participants. Periodically, comparing actual outcomes to projections should test assumptions. The resultant feedback can then be used to adjust our assumptions and treatment models.
Some of the basic clinical parameters that are typically collected include:
- DSM-IV diagnosis, all axes
- Presenting problem
- Level of functioning
- Status of goal attainment
- Recidivism
- Retention (completion of treatment)
- Client and family satisfaction measures
The treatment parameters that should be monitored include:
- Program
- Intervention
- Clinician
- Level of intensity of service
- Length of stay/treatment
- Number of services
- Treatment goals
- Treatment objectives
Some of the demographic variables that are typically collected include:
- Age
- Gender
- Ethnicity
- Education
Some of the simpler cost variables that are collected include:
- Calculate rates for services by case, by episode of care, by covered population
- Calculate write-downs for discounted fees by payer
- Prepare benefits report that lists clients whose benefits are about to run out
- Prepare uncovered services report that lists clients who received services that were not covered
- Budget variance reports that list budgeted amount used and remaining dollars by client and program
- Capitation analysis report that compares the cost of services for a particular plan with the budget
Despite the challenges of collecting data in a clinical environment, the use of outcome monitoring is becoming increasingly important for evaluating treatment protocols, achieving accreditation and winning managed care contracts.
To gain the most benefit from these data they need to be collected in a manner which facilitates their aggregation and analysis. Thus the quantitative analysis of patient outcome data requires the use of standardized treatment protocols, data collection procedures, and coded or numerical responses. The data, if not initially collected in the computer, should be summarized and entered into the patient record in a way that permits its association of other patient related variables. This may be accomplished by the use of a sophisticated ad hoc report writer or by downloading a file into a commercial statistical package like SPSS or SAS.
About The Authors:
Bruce Johnson, M.S., is President of Johnson Consulting Services, Inc., an information management consulting firm that specializes in working with healthcare, social service and managed care organizations. He can be reached at (800) 988-0934, www.jcsconsultants.com or by e-mail at jcs@eos.net. Mr. Schafer is a clinical records and operations management consultant. He specializes in working with managed care, behvavioral healthcare and child welfare organizations. He can be reached at (800) 661-2435, www.schaferconsulting.com or by e-mail at steve@schaferconsulting.com.
|
|
 |
