BACKGROUND: Lithium (Li) is a “gold standard” treatment for bipolar disorder and possibly for preventing suicide. However, Li also has unique hazards due to its narrow therapeutic index. The suicide prevention potential, toxicity risk, and monitoring of Li has become an intense focus of two VA offices, OMHSP and PBM. This study builds on strong partnerships with OMHSP and PBM to develop an innovative near-real time system to inform Li prescribers of events that sharply raise the risk of Li toxicity and of needed monitoring. METHODS: This 1-year pilot study will develop a “Lithium Support System” (or “LiSS”) using near real-time CDW data to track and notify prescribers of both upcoming and past due monitoring and of the occurrence of clinical events that are powerful risk factors for Li toxicity. These clinical events are the initiation of interacting medications by other providers, certain medical illnesses (kidney problems, infections, dehydration, hyponatremia), and dose increases. Perhaps surprisingly, the VA’s computerized patient record system (CPRS) does not provide notifications currently to Li prescribers in any of these circumstances. Our strong partnership with OMHSP is evidenced by the fact that they have committed to provide us with code they have already used to track overdue Li monitoring. The system will provide daily reports from CDW which will be reviewed by a central evaluator (to prevent spurious notifications), with notifications then sent to prescribers via encrypted email (and sometimes voicemail). (Email, rather than CPRS, is used in this pilot to reduce “alarm fatigue”, facilitate tracking when notifications are read and prescribers’ immediate impressions to them, and to permit guidance about appropriate clinical responses). The system will be test implemented at 2 sites (Bedford and Boston VAMCs). We will evaluate process data (e.g., number of each notification type generated, central evaluator time required, provider response times) and chart data regarding timing and types of clinical responses (e.g., educate patient about toxicity symptoms, hold Li, decrease dose, switch medications). Survey and interview data will be used to assess the usability and acceptability of the system. Nationwide data analyses will seek to identify other predictors of Li toxicity to add to a “Phase 2” enhancement of the system introduced in the study’s final months. Nationwide data will help estimate the toxicity episodes that might be averted with broad implementation of the initial or expanded system, and examine clustering of toxicity events. SIGNIFICANCE: Importance: Li toxicity is one of the major health concerns and perceived barriers to the use of this highly valuable medication. Timeliness: Development of this system is urgent, given VA’s intense interest in reducing Li toxicity episodes and the fact Li use may increase as a result of OMHSP’s efforts or a large-scale Li suicide prevention trial approaching completion. Presumed Cost-effectiveness: This system, like the other centrally-operated tools using near-real-time CDW data being developed, should improve care across the VA, especially at low-performing sites, while operating in a manner likely to be relatively low cost and cost- effective. Unique Scope: The system’s ability to go beyond simple monitoring prompts to also warn about events that may lead to lithium toxicity is truly novel. Flexibility: The system, once developed, will be easily customizable, both in the range of the notifications and the specific information included in notifications. The wide range of process, clinical and usability data we will gather will facilitate decision-making about subsequent implementation. Broader implementation might range from larger research-based implementations to largely program-led implementations of all or part of the system. Eventually it is hoped that the whole system, or even a further expansion (e.g. providing information about patient lapses in refilling Li), could be implemented nationally. This might prove especially helpful if Li becomes a centerpiece of VA suicide prevention efforts. The result of this research will be an urgently-needed, very timely, and highly cost-effective study with extremely strong operational partnerships that may ultimately help improve care for thousands of Veterans.
External Links for this Project
Grant Number: I21HX002535-01A1
Dimensions for VA
None at this time.
Mental, Cognitive and Behavioral Disorders
Treatment - Observational, TRL - Applied/Translational
Bipolar Disorder, Medication Management, Suicide
None at this time.