At the University of Kentucky Children’s Hospital, when expanding their Neonatal Intensive Care Unit, the owner chose to explore a plan using combined renovation and expansion space. Due to existing floor configurations, all patient rooms were designed without windows. In addition, a procedural change from dark, shared rooms to single rooms emulating a day/night cycle were planned. As a result, the lighting was a critical element in the success of the project.
Project design parameters included providing 600 lux at the isolette, indirect lighting with “no ledges” per UK design standards and a 0-10-V control system. Due to the infancy of tunable white LED systems at the time, a simple 7 a.m. to 7 p.m. schedule changing between 3000K and 5000K was planned. Color and intensity did not fade over the course of the day but was programmed to transition quickly at set intervals in the morning, evening and nighttime hours.
Despite the intended simplicity, as physical construction ended and the control sequence of operation was being finalized, the complexity with which the control system was required to operate and respond to users in the space became a reality. If the lights were turned off during a procedure, would the control system be able to turn them back on to the correct programmed intensity and color? What color and intensity would lights display in the middle of the night during nursing checks? The complexity of the control system design, exacerbated by the intricacies of the healthcare environment, limitations of LED driver dimming range and required control protocol, were an element of “circadian design” not fully comprehended at the beginning of the project. Painstaking consideration for each touch of a button was required to ensure successful operation of the system.
With construction complete and systems commissioned, it is typical for the design team to withdraw from the project, hoping the owner is satisfied and all systems will continue to run as intended. However, with this project, we began wondering how the system would really be used. Would the transition from the existing dark environment to a bright new one be so extreme that staff would turn the lights off each day? If turned off for a procedure, would staff remember to turn the lights back on correctly? When asking those questions, we learned that the control system could record and track each button setting through the course of the day. The hospital was willing to provide the data recorded to Pacific Northwest National Labs (PNNL) for their study and evaluation. (See technical paper “Lighting System Control Data to Improve Design and Operation: Tunable Lighting System Data from NICU Patient Rooms” in the May 2022 edition of LEUKOS). The alignment of these opportunities has provided valuable post-occupancy insight to our design team.
Using the data collected in five-minute intervals, from five patient rooms, over the course of 25 weeks, PNNL evaluated the basic system programming in comparison to daily use. The results were eye-opening. While the system was intended to provide a consistent scheduled pattern each day, the data enabled us to compare programmed settings vs. daily manual adjustments. It was evident that when the lighting was turned off during the day for special procedures, there were many instances when it was not turned back on to the programmed state. For nighttime hours, when light was required for patient checks, a preset was created to provide low illumination at the bed side. However, there were no instances where the preset was used as programmed. Instead, there was a 50/50 split between the exam and custom modes, with the nighttime preset light level being increased 100% of the time it was used. This information not only indicates that the preset level was too low, but also presents the concern that exam lighting turned on in the middle of the night would create sleep disruption, mitigating the effects of the cycled system overall.
Through reflection on the project design and post-occupancy evaluation, in combination with current product availability, the design team has been presented with several opportunities for refinement on future projects. Key attributes for future consideration of lighting products include LED drivers that have a full dimming range to provide smooth lighting transitions, non-proprietary LED board/control combinations, and the ability to graphically select color temperature from the control system to simplify programming. Significant technological advancements since project completion offer a variety of new products—but each must still contain these critical elements.
During review of the manual override issue, the design team discussed several items that could be enhanced. For instance:
- To minimize high-level nighttime overrides, the preset default ON level could be programmed slightly higher so that nursing staff does not need to increase the illumination.
- To reduce the number of times that FULL ON exam lighting is required, the raise/lower speed at which the overrides respond to manual changes could be increased.
- To provide adequate nighttime illumination at the primary task locations with minimal spill into the remainder of the room, validation of primary task locations in combination with specifying luminaires using improved optics could more effectively reduce the need for a general high-level override during nighttime patient checks.
By minimizing the high-light-level overrides, we can reduce sleep disruption. It is understood, however, that overrides for emergency or procedural reasons are required. Upon data review, it was evident that after a system override, the lighting was frequently not returned to its programmed state. As a result, the lighting often remained off through a portion of the morning or remained in exam mode during the day. We understood this may be an issue with the control system installed, but the data made it clear that the system had limitations.
Selection of systems and other elements for products moving forward must be considered. For example, improved button preset labeling could improve staff understanding, or better training could be provided. Or, rather than being dependent on staff for a program reset, an automatic time-out would be more effective, returning the lighting to the pre-programmed color, spectrum or intensity.
Through the design and completion of each project, we as designers are continually adapting to new technologies and project requirements, fostering the evolution of design and product development. The UK NICU project has informed elements during design of the recently constructed Cincinnati Children’s Hospital Medical Center (CCHMC) NICU and is continuing to inform our approach to the current design of the Houston Methodist Hospital Centennial Tower.
Although the system at Cincinnati Children’s is entirely different from the UK NICU and highly customized in response to the research intended, there are elements that the previous design informed. In each patient room, a manual override of the programmed circadian system was initially going to be excluded based on research team requests. Upon sharing our previous practice and experience, design changes were made and a manual override for staff and families was included. The override is only for ON/OFF but it returns automatically to preset programming, ensuring the day/night cycle is repeated each day. This is achievable due to the complexity of the customized control system that tracks the intended color, spectrum and intensity throughout the day, ensuring that manual overrides do not interrupt the underlying programming. Although this operation appears straightforward, there is significant complexity required. Through advances in technology, this control is becoming more available in standard product, but it is critical to confirm the operability of any proposed control system prior to making a final selection.
Looking ahead to the Houston Methodist Centennial Tower that is currently in final design, previous experiences are informing design decisions in other ways. Although there is no NICU, the raise/lower timing of the manual overrides within the patient room low-voltage control system are being modified to increase response speed. With the intent to minimize sleep disruption, the goal for overrides to respond more quickly reduces the need for nursing staff to turn more lights on in the middle of the night. With the knowledge gained from the UK NICU project, in combination with questioning current product standards, this simple change will improve the patient/ staff experience. Elsewhere in the building—to improve staff well-being and satisfaction in select windowless areas where daylight is inaccessible during the day due to the large floor plate—we have proposed a simplified lighting system designed for circadian support. Based on understanding of the highly complex systems in the previous two facilities, this simplified approach reduces the additional cost often associated with these systems. By using a targeted spectrum LED board with a single 0-10-V channel of control, the control system specified for other areas of the building is also able to operate this one. Despite the simplicity intended, it is well understood that conversations are still required to understand staff shift length, provide suitable manual overrides and appropriately educate the staff.
Technological advancements in the lighting industry enable the design of very complex lighting systems. However, as designers, we must continue to ask critical questions during system and product selection processes to ensure efficient and effective system operation. The rare insight from the above examples provides a deep data review, allowing the design team to implement future improvements and inform similar designs—ultimately leading to increased staff satisfaction and better patient outcomes.