TCC’s Department of Data Analytics
By: Dr. Eli Sprague
A year in review, 2022
“Information is the oil of the 21st century, and analytics is the combustion engine.”—Peter Sondergaard, Senior Vice President and Global Head of Research at Gartner, Inc.
Data Analytics has arrived to TCC! Beginning in January 2022, Dr. Eli Sprague met with CEO Andy Albrecht and Senior director Max Liles to explore options of establishing outcomes related to Alcohol and or Drug treatment (AOD) and Mental Health (MH) services. The primary task consisted of assessing the agency’s performance in their provision of services. As work began to build up, the team required more hands-on deck to mine and process the data. Holley Hart , an experienced treatment engagement specialist with extensive knowledge in excel and other administrative tools, was brought on board as a data specialist to help provide valuable insight. In addition, Destry Lowe applied for the position and carried with him governmental administrative acumen with experience in data visualization tools such as Tableau. The team set out to understand clinical outcomes and to assess for areas in need of improvement. The following is a review of the Department of Data Analytics (DDA) annual review.
Voluntary Early Termination
Voluntary Early Termination, colloquially known as ‘Against Staff Advice” (ASA) was one of the first items in the DDAs agenda. The team reviewed data from NextGen and Survey Monkey. Initial findings revealed excessive quantities of patient-ASAs within a three-year period. Accordingly, 27.8% of all admissions left ASA during their treatment stay. It was revealed that both length of time and time of year played a significant role in determining a person’s ASA. For example, Women were more likely to leave ASA during colder months, especially around November. In fact, 78% more women left in November than did their male counterparts. The team’s initial speculation involves the holiday’s being a major trigger for women regarding societal mother-expectations to care for their families during the holidays. However, these speculations have not been validated by follow up studies.
In addition, ASA rates were linked to length of stay (LOS) as the longer patients remained in treatment, the less likely they were to leave treatment against staff advice. Accordingly, for every 1-2 weeks of time a patient remained in treatment, their likelihood of leaving ASA decreased by 30%. The implications for these outcomes are instrumental in determining strategic interventions to reduce voluntary early termination of services. For example, increased efforts to provide more attention to newcomers will play a key role in helping them to remain in treatment (Efforts may include but are not limited to creating two treatment tracks for ease of integration). This would involve creating a slightly different experience for newcomers who may require less intensive treatment interventions.
As mentioned, one of the primary tasks of the Department of Data Analytics was to examine outcomes of patients entering treatment. To this end, a series of metrics were used to monitor symptom severity within 30-day time intervals. These data would demonstrate observable trends about patient symptom expression throughout their treatment. Accordingly, 6 scales were used to assess symptom expression including DUDIT (drug-use), AUDIT (alcohol-use), PHQ-9 (depression), GAD-7 (anxiety), PCL-5 (PTSD), and SWLS (life-satisfaction).
To continue to monitor patients and track their monthly survey responses, an excel spreadsheet was created with color coded cells to capture date of future surveys (See image below). Under the status column, red indicates the patient is past their 30 days expiration marker and requires a survey immediately. Yellow indicates that within 5 days the patient will require a completed a survey. Green notes that a patient is within an acceptable amount of time before their next survey is due for completion.
To increase efficiency, emails are sent to providers en mass via an automated emailing system which is linked to the survey tracker sheet. These emails are sent on the basis of their current survey status (i.e. past due, almost due). In addition to an email which includes the survey link, providers are also reminded via their Next Gen Clinical schedule to complete the survey with their patient. This method increases the likelihood that providers will see their obligation to complete the survey.
The department of data analytics (DDA) began analyzing the information to identify any patterns. The initial findings were limited with a small sample size (n = 3). The team observed an overall decrease in symptoms of anxiety and depression throughout the treatment. For example, within the first 30 days of treatment, the average GAD-7 score—used for assessing anxiety severity—was 10 (interpreted as moderate levels of anxiety). Over a period of 90 days these symptoms decreased by 62% dropping to an average score of just 5. In addition, Patients experienced average score of PHQ-9—used to assess depression severity— of 8 (indicative of mild depression). Like anxiety, patients experienced a sharp reduction in depressive symptoms of approximately 66% within the first 90 days of treatment. With over half a reduction in anxiety and depression, it can be presumed that treatment is more than likely having some impact on patients’ mental health.
Moreover, PCL-5 scores—used to assess PTSD symptom severity—demonstrated a net decrease in average scores over a period of 90 days with an average score of 27 at the beginning of treatment followed by a reduction to an average score of 7 within the first 90 days of treatment. These scores demonstrate a net decrease of 74%. Although the onset scores are relatively mild, notwithstanding, these numbers demonstrate a sharp decrease in symptom expression which can interpreted as improvement of the patient’s mental and emotional well-being.
In addition to program outcomes, The Counseling Center requested assistance from the Department of Data Analytics to aid in applying for the Appalachian Community Grant Program (ACGP) which was enacted by Governor DeWine in June 2022. The purpose of ACGP included the direct allocation of funds for revitalization of Appalachian communities in Ohio. A total of $500 million dollars was made available to 32 Appalachian counties. The decision to proceed with the first round of funding was made by the board to apply for the Appalachian Development Grant which included a total of $50 million dollars.
Destry Lowe, Master of Public Administration, utilized his knowledge of public administration and employment experience to spearhead the project narrative. In so doing, the team was able to gather resources to complete the official grant application. These resources included aforementioned data on clinical outcomes in conjunction with internal employment descriptive outputs from a survey conducted on TCC Alumni. In addition, the team created a heat map which consisted of all the counties in the state of Ohio which TCC receives patients from.
The summation of the DDA’s collective efforts for the application reveals interesting occupational trends for the success center and the TCC alumni. For example, it was revealed that 2021, 46 vocational courses were provided with a total of 281 participants in attendance. In 2022, we saw an increase in vocational courses (59) and a total of 459 vocational participants in 2022. When using the growth function in excel, (a tool that predicts future amounts based on previous trends) it was predicted that by the year 2026 there would be 159 classes offered with over 3,000 participants.
Conclusion and Future Implications
Within the first year of operation, the Department of Data Analytics (DDA) provided a summarization of the need for treatment within the Scioto Co. Region which was then used in the Appalachian Community Grant Program (ACGP) application. The department’s findings were based upon data from the Ohio Department of Health and their 2020 Ohio Drug Overdose Report. After conducting a regression analysis, an exponential growth was detected in drug overdoses in Scioto Co. This means that drug overdoses continued to increase 20% each year. Based upon previous trends, the growth function in Excel was again used finding that if trends continue, there will be an expected overdose rate of 320 overdoses per year by 2027.
The available data demonstrates both the need and potential for continued efforts in data analytics. Our models demonstrate that addiction and drug overdoses will likely continue to intensify with each growing year. Adding to this, the supply of dangerous drugs such as fentanyl is likely to increase. According to Lim et al. (2022), projected fentanyl overdoses are expected to occur between 543,000 and 842,000 between the years of 2020 to 2032. Increased efforts in data analytics are required to ensure better understanding of substance and the variables that both help and hinder patient’s recovery.
The year 2022 marked a turning point for the Counseling Center as it decided to apply resources to data analytics. The year 2023 includes plans to take their initial progress and continue to expand on it. Accordingly, DDA will begin using their skills in other departments in need of analytics. For example, DDA has begun working with human resources to develop a measure of employees and their understanding of ethics. This tool will be used to assess future employee’s capacity to take on the role of working direct care. Beta testing of this measurement tool is set to begin on January 16, 2022, within the success center. After collection of data, the team will begin to perform tests such as Cronbach’s alpha and confirmatory factor analysis (CFA) to assess and refine the instrument for future implementation. After significant data is acquired for one year of the instrument’s use, analytical methods such as canonical correlation will used to asses’ correlations between scores and employees length of employment.
In addition, the team plans to expand the capabilities of software packages by using analytical tools including, but not limited, Tableau, R, and Python. Moreover, the team plans on expanding the capabilities of the survey instrument platform known as Qualtrics to begin using Application Programming Interfaces APIs to increase efficiency of data collection which will in turn provide clinical data specialist with the ability to rapidly acquire real time reports of patients and their current mental and emotional functioning.
Lastly, the team plans on using their skills toward future grant funding acquisition efforts. One of the team members, Destry Lowe has extensive skills in writing grants and navigating bureaucratic structures and will be a significant asset in acquiring future funding. Moreover, the teams’ skills will be useful in negotiations with managed care officers as the team will be able to verity, quantitively the mental and emotional improvements of patients over the course of treatment. As a result, DDA will continue to bring value and prestige to Counseling Center which will in turn help the community it intends to serve.