General Lessons Learned
A number of lessons were learned during the course of the project. First, it took a number of months to obtain the necessary approvals to conduct the project, and to obtain the project data extracts. Approvals included I/T Institutional Review Board approvals and Tribal council approvals; data use agreements also had to be enacted. It is important to include a realistic timeframe for such vital project steps when working in partnership with I/T health programs.
Second, collaboration with the project sites was essential for the implementation of this project. The collaboration included development of project measures and reports that would provide actionable information on the health status of sites’ populations, provision of services, and health spending. Each site was provided detailed project site reports that included data similar to those presented in this report, and which could be used to identify strategies for enhancing services. When possible, it may be useful to compensate sites for their personnel time allocated for similar efforts in the future.
Finally, project sites may require additional resources to fully implement programs and policies identified to enhance services. Thus, it may be useful to identify options for increasing available resources (e.g., reimbursement for services). Implementation of the Patient Protection and Affordable Care Act will provide AI/ANs increased opportunities to obtain health coverage, and I/T providers increased opportunities to obtain reimbursement for provided services,11 specifically those aimed at preventing the need for more specialized and costly services.
Suggestions for Modifications to the National Data Warehouse
The NDW includes over 30 data tables with electronic data for I/T health programs. The data may be used to conduct a number of valuable analyses to inform the provision of services, as well as policies and programs developed to enhance services. Based on our analyses of these data, we provide recommendations for enhancing specific NDW measures and adding others.
Service Utilization Measures
Clinical Pharmacy Services
We used provider codes, clinic codes, and CPT codes to identify this type of service. Based on consultation with IHS pharmacy experts, there is general agreement that, using available data, we were not able to identify all provided clinical pharmacy services. Identification of these services, as well as ECM services in general, deserves additional consideration.
Use of Prescribed Medications
For each medication, the NDW includes data on service date and quantity dispensed. In order to more fully understand pharmacy use, it is important to have a measure of the number of days the medication was provided for (i.e., days’ supply). This data element is available in the data stored in most project site data systems.
CHS Utilization and Payments
Analyses of NDW CHS data and CHS data obtained directly from the IHS fiscal intermediary should be conducted to ensure that the NDW includes up-to-date CHS information and all relevant data variables from the fiscal intermediary. In addition, IHS may consider modifying the place of service codes included in the fiscal intermediary data to include more detailed information on service type. For example, Medicare place of service codes may be used.
Software exists to identify hospital readmissions that are deemed preventable and to analyze lengths of stay and clinical severity. The software may be used to develop additional measures of inpatient service use to monitor utilization and identify opportunities to enhance services.
Software exists to classify ED visits into clinical categories to be used in assessing disease severity and use of ED services for treatment that may be provided more efficiently in other settings.
Project sites use this clinic code in various ways, due to differences in the provision of services. It may be useful to create a standard definition.
Cholesterol Laboratory Test Values
On average, LDL cholesterol test values were missing for half of all adults with diabetes. Due to the importance of measuring cholesterol among persons with diabetes and CVD, it may be worthwhile asking project sites to report some laboratory test values, such as LDL test values, in specific data locations on their local computer systems so that the data may be included in the NDW.
Measures of Kidney Function
Across I/T health programs, a number of tests are used to measure kidney function. Due to morbidity associated with kidney disease, it may be worthwhile to identify a method for including measures related to the type of test and the test outcome in data fields in the NDW. Currently, some outcomes are provided in text fields that are not generally useful when conducting analyses across project sites.
Suggestions for Increasing Access to and Use of Existing Electronic Data
It is possible that data analysts and statisticians who work within I/T health programs do not access the NDW data for analyses due to the number of data tables and the complexity involved in linking data from different tables for many analyses. To increase use of existing NDW data for planning, quality improvement, and other purposes, we suggest creating user-friendly data files, as we did in this project, from the NDW and other existing data sources. Chapter 3 provides information on the data files in the project’s data infrastructure. Not only did we use the infrastructure’s encounter and person summary data files to conduct the analyses in this report, we provided 7 of the 14 project sites copies of the data files for persons living in their project sites. During the July 2012 Project Data and SPSS Training Program, we trained 11 I/T personnel on use of these data. The program participants reported the training program was valuable, and that having access to detailed information on patients with diabetes would inform program planning and evaluation.
We recommend that the project data be made available for use within various IHS divisions. As described above, the data may be enhanced by including other existing data and by creating additional measures through use of other software programs. Furthermore, the data infrastructure could be expanded to include data from other project sites. It is possible to update the analytic data files quarterly to provide more current information than that available with annual updates (with the exception of the cost estimates, which may be updated annually).
Expertise in health services research (e.g., economics, software programs) and advanced technology (e.g., computers with sufficient processing speed, data storage capacity, and security) are required to update the data infrastructure files efficiently; therefore, we recommend that outside contractors be used to develop the quarterly and annual updates. In addition, the contractors could provide online training to I/T health program personnel on use of the data mart. It may be possible to form a working collaborative of I/T health programs and the participating project sites to identify meaningful uses of these data and to prioritize enhancements to the data mart.