Summary
This multidisciplinary study is focused on the potentially beneficial “non-visual” effects of lighting in the clinical environment, to improve clinician wellness and performance, and advance patient safety.
The hypothesis of this study was that clinician wellness and performance in the execution of clinical procedures at the emergency department (ED) could be improved through controlled, indirect, “blue-regulated”, full visible spectrum, tunable, solid state, “white” lighting, compared to prevalent fluorescent lighting conditions.
To conduct our inquiry, we performed a randomized AB/BA crossover experimental study with ten (10) ED clinicians that executed clinical procedures, under two lighting conditions, in a realistic ED clinical setting. The experiments took place at the Emergency Department Simulation Teaching and Research Center (ED-STAR) at Mount Sinai Hospital in New York City, NY, USA, with a high fidelity human patient simulator.
The main observed outcomes of this study were: a significant reduction in clinician sleepiness perception (Karolinska Sleepiness Scale, “KSS”, -15.94%, p = 0.022), a significant reduction in workload perception (NASA-TLX, -21.87%, p = 0.009), a reduction in clinical procedures execution time (-21.04%), and a reduction in the occurrence of medical error. The experimental condition (78,000 K) was also preferred by the clinicians compared to the prevalent fluorescent lighting (4000 K), favoring the translatability from the simulation setting to the clinical environment.
We concluded that controlled lighting environments may contribute to improve clinician wellness and performance, reduce medical error, and improve patient safety. Therefore, lighting should be considered a critical factor in the design and operation of health care facilities, far beyond energy savings.
Abstract
Short wavelength (“blue”) light is known to mediate “non-visual” effects of light in humans [81,16]. These effects that go beyond the pure “visual” function can affect human wellness and performance, as it has been reported in previous scientific research in laboratory, office, education, clinical and aero-space setups [54,47,29,60,80, 4,5,6,83,44,17,45,11]. In healthcare research, lighting has been recognized in the fields of human factors, ergonomics, and systems engineering, as an environmental factor that can affect the quality of the delivery of care; in particular, clinician wellness and performance, and the occurrence of medical error. The aim and novelty of this research is to study the potentially beneficial “non-visual” effects of lighting in the clinical environment to advance patient safety.
The hypothesis of this study was that clinician wellness and performance in the execution of clinical procedures at the emergency department (ED) can be improved through controlled, indirect, “blue-regulated”, full visible spectrum, tunable, solid state, “white” lighting, compared to prevalent fluorescent lighting conditions.
To conduct our inquiry, we performed a randomized AB/BA (2×2) crossover experimental lighting study with ED clinicians that executed clinical procedures, in a realistic ED clinical setting, under two lighting conditions. We used the existing fluorescent lighting as control condition. To provide the appropriate experimental lighting condition, we developed a novel multichannel solid state lighting (SSL) system for precise control and assessment of the light spectrum, with specific emphasis in the short wavelength spectral area.
The results of this study suggest that it is possible that indirect, “blue-enriched”, full visible spectrum, tunable, solid state, “white” lighting, can contribute to reduce clinician sleepiness and workload perceptions, might reduce clinical procedures execution time, and might reduce the occurrence of medical error (compared to prevalent fluorescent lighting conditions).
Background
Health care is a critical burden for modern societies, and patient safety has become a main concern in the last decades [48]. The study of the (internal) environment of care is emerging and has not been properly addressed, even if research in human factors and ergonomics points it out as a critical factor [12,13,14]. In addition, lighting, as a key factor of the environment (more specifically, of the “exposome” [86], has not been paid enough attention except for the considerations of “adequate lighting levels” for the visual task and for visual performance [13]. Improving the environment of care can have a positive impact in the quality of care and therefore in patient safety [68,42]. Other studies propose that enhancing clinician wellness might contribute to this goal [84,69]. And this is also a goal of our research, to see if through controlled lighting interventions clinician performance and clinician wellness can be improved with the final objective of having a safer delivery of care, and improvement in patient safety.
The recent adoption and development of simulation centers as part of the medical training and research provides a very appropriate setting to conduct research without the risk of harm to patients and with a high level of translatability to the real clinical setting [34,32]. For these reasons, we have conducted our research at the ED simulation center, to provide a realistic reproduction of the clinical environment. We know that the human “visual” system has, at least, a dual function, the “visual” and the “non-visual” [50,70,23]. Besides the known pure visual effects (mainly visual task and aesthetics), and the non-visual or non-image-forming (NIF) effects of light and lighting on circadian rhythms that can affect, entrain and regulate sleep patterns, disorders and disturbances (such as shift work and jet-lag) [35], research is developing in the fields of effects on human mood, cognition, performance and wellness and wellbeing [82,18]. Although there has been extensive research and development of lighting for health care delivery, mostly for the patient [80,4,5,6,83,44,17,45,11], there is no prior interventional research that specifically addresses the NIF effects of the lighting environment in the ED clinician performance and wellness in relationship with patient safety, what represents one of the main contributions of this study. The field of research on NIF effects of light and lighting is emerging, and has several limitations that will need further research. Appropriate research in the NIF effects on lighting has only been possible since the recent discovery in 2002 of the ipRGC [39,7], the photoreceptors that mainly mediate these effects, and their presumed “action spectra” (related to the photopigment melanopsin [8,78], also recently discovered in the late nineties by Ignacio Provencio, Ph.D [63,64,65], a reason why some researchers call the ipRGC Ignacio-Provencio-RGC). While the precise underlying neural pathways and systems and associated “pharmacodynamics” remain unclear and will need further research from multidisciplinary approaches (such as architecture, design, engineering, physics, medicine, biology, cognitive sciences and psychology), it is known that these ipRGC photoreceptors innervate into different areas of the brain far beyond the well-known visual pathways [24].
There is a general lack of methodology in lighting research, particularly in NIF. Replication and reproducibility are basic pillars in the scientific method that must be addressed in this field. Uniformity is needed in lighting studies to allow repetition and replication and conduct proper metaanalysis [66]. Methodology becomes essential for lighting studies in the field of NIF effects. The following are proposed factors that have to be carefully considered [66,53], as subjects’ psycho-physiology plays a major role: participant’s chronotype (morningness/eveningness); sleep schedule of the participants during previous days, circadian time of day (CT) for the participant (time since wake-up) and other stimuli influence such as caffeine intake (in the form of coffee or other caffeinated beverages and/or supplements), Indoor Environmental Quality (IEQ), with factors such as room temperature, relative humidity and carbon dioxide (CO2) and noise; appropriate method of light measurement, preferably based on spectral power distribution (SPD). Control conditions are also critical for the estimation of the effects beyond the classical “statistical significance” research drivers. Using darkness [71] or unrealistic environments as controls is not ethical. These are factors that need to be properly considered in future studies. We have seen the same lack of rigorous methodology in the review of most of the studies referenced related to the effects of light and lighting in the health care environment.
One of the main limitations in current research is the paradigm, and prevailing mindset, of photometry and colorimetry in the lighting community. Light and lighting are mainly measured, and understood, in terms of illuminance (unit lux) and color appearance (correlated color temperature, “CCT”, unit Kelvin) [27]. The “action spectra” of the ipRGC, characterized now in relationship to melatonin suppression [8,78], a point that is beginning to be questioned [67,87], is totally different from the action spectra of the classical visual photoreceptors that are the foundation for photometry and colorimetry. New metrics, such as the melanopic function (and the associated melanopic lux), or the “melanopic equivalent daylight (D65) illuminance” to be introduced in late 2018 by CIE JTC-09, have been proposed for the measurement of light at the eye/cornea level in laboratory studies [53], but there is no common standard for measuring the field of view (FOV) and the whole lighting environment, particularly with spectral resolution (SPD). Future advancement in this field will require proper assessment of the directionality and of the whole lighting environment to better characterize the dynamic stimuli for the different space stakeholders “competing demands”, such as patient vs. caregiver in the patient room or the emergency department, and the surgeon vs. anesthesiologists vs. nurses in the operation room.
Technology availability by lighting researchers has constituted an additional limitation. Most of the previous studies in real or realistic environments have been done by nonexperts in lighting with commercial light sources and fixtures designed for the visual-task function and performance. Little attention has been paid to these factors in lighting design improvement, limiting interventions (control vs. experimental setting) to lamp substitution or fixture retrofitting. Solid state lighting, and particularly LED (light emitting diode) facilitates a radical change in this limitation, particularly the “full spectrum tunable” (not the inappropriate “white-tunable”) systems that will allow targeting specific SPDs.
Due to this current limitation, and to our experience in the development and commissioning of lighting systems, we decided to set up in our research environment an innovative lighting system that provided enough flexibility for our research at an affordable budget. We describe this in more detail in the methods section.
Research Questions, Hypothesis and Objectives
Research questions
The main research question of this study was: “to know if there is a difference in clinician wellness and performance during the execution of clinical procedures at the ED Simulation Center due to the lighting environment”. Performance was measured in qualitative terms (medical evaluation), and in differences in execution time, together with subjective workload and sleepiness/alertness metrics for wellness evaluation. Medical error occurrence was also surveilled.
We were on an exploratory research path and this was a screening experiment. Therefore we were not interested at this stage in knowing the optimum spectrum or the most efficacious system. If we obtained satisfactory proof of concept findings, as it happened, this matter would open other research oriented to optimization of our research question, to estimate the best control variables (factors) and to get the optimum values for them.
There are two additional research questions of interest:
• What is the effect of “blue enriched” indirect lighting in lighting
quality perception?
• “Blue-enriched” lighting, associated with high CCT, has generally not
been liked by previous study participants. Is our lighting environment,
diffused lighting, going to be accepted by our participants?
These secondary questions are critical for the translateability of the results into the real clinical environment.
Hypothesis
The research hypothesis was: “the clinician wellness and performance outcomes will be significantly improved with the indirect, “blue-enriched”, full visible spectrum, solid state, “white” lighting (experimental treatment condition) compared to the existing fluorescent lighting (control treatment condition).” Based on our previous experience and in the literature research, already discussed in the backgrounds section, we presume that “blueenriched” lighting can be adequate to be part of our research. And we also know that this particular region of the spectrum can be difficult for our visual system if not delivered properly. That is why we chose to research with an indirect delivery of the lighting conditions, with no direct exposure of the light sources to the participants. This also minimizes the controversial “Blue Light Hazard” (BLH) risks [62].
Objective
The objective of the study was “to investigate the efficiency of indirect, “blue-enriched”, full visible spectrum, solid state, “white” lighting in clinician wellness and performance, and patient safety, during the execution of a single-clinician clinical procedure, “chest-tube”, in the Emergency Department simulation lab.” The first procedure was taken as a “warm-up”.
Methods
We will discuss here the experimental design (AB/BA crossover, within subjects), the research tools (with a special emphasis in the lighting intervention and the lighting system), the population, and the statistical model for the study (that has to be properly defined before running the tests)[i].
The whole experimental flowchart is shown in figure1. There are three main processes, the experimental design and IRB (Internal Review Board), the experimental runs, and the data analysis, though we will not explore them here in detail.
Figure 1: ED-STAR experimental methods flowchart
Experimental design
We conceived, designed and conducted “proof of concept” experimental research. We had two lighting conditions (control and experimental), and we randomized our sample of participants (real ED clinicians) into two groups, sequences AB and BA. Group AB performed the test (two sequenced clinical procedures “airway” + “chest-tube”) first under control lighting condition (fluorescent) and then under experimental (“blueenriched” LED), and BA group did the reverse order/sequence. Participants took a “washout” period of at least one full week (7 days) between the treatments/ conditions (tests). The process flow is shown graphically in Figure 2.
Figure 2: Experimental design AB/BA crossover flow
Research tools
For the experimental lighting condition (treatment), we developed our own instruments and systems and used as the control the existing lighting system (fluorescents). This system will be described in the next subsection.
For the experimental instruments, we used validated surveys and questionnaires: the “Karolinska Sleepiness Scale” (KSS) for subjective assessment of sleepiness/alertness [2,46,75,41], the “NASA-TLX” (“TLX”: “Task Load Index) for subjective workload perception [37,38,49], a modified version of the “ASHRAE IEQ survey” [88] for subjective assessment of the environment, and a “Lighting Perception Survey” for the subjective perception of lighting, based on the initial work by Prof. Peter R. Boyce [30], that we updated to gather additional information. We used custom ED-STAR tables for clinician performance and time evaluation.
Lighting intervention and lighting system
At the time of the study, there were no commercial options to provide illumination levels of a room such as the ED-STAR “simlab” for the purposes of the experimental framework proposed here, with full control over the visible spectrum (SPD). The only existing systems that were able to provide room level “color tuneability” (different from SPD control) were the Philips “Healwell” and the Telelumen “Penta” that provide color (perceptual) tuneability, but not detailed spectral tuneability. This, as discussed, is particularly relevant on the “blue” region of the visible spectrum (short wavelengths) that is critical mediators of the “non-visual” effects of light. Besides the SPD limitation, both systems are based on direct ceiling light emission, contrary to the IESRP-29 recommended practice for healthcare facilities, which prioritizes indirect lighting for spaces that have interactions with patients. For these reasons, we proceeded to build a custom system at the “simlab” which provided full SPD controllability with an indirect lighting strategy.
Figure 3 shows the “simlab” room as it was in May 2015 prior to proceeding with Mount Sinai facilities to retrofit the room for this research. The room was originally illuminated by four recessed ceiling troffers (2’×2′ fixtures). The walls were painted in an unspecified “off-white” with the head wall painted (arbitrarily) in a “beige” color.
Figure 3: ED-STAR room before the lighting intervention
The first intervention was to paint all the walls of the room with a “high reflective white” (Light Reflectance Value, LRV = 94%) paint to make the room an integral part of the lighting system. We also painted two ceiling tiles with the same paint, as they were going to be working as primary reflectors of the light engines that were providing full spectrum tuneability. This also provided maximum spectral fidelity of the reflected light and minimum absorption by the surfaces (a factor that should be considered also for energy efficiency). Figure 4 shows the final configuration of the system as conceived for the experimental condition with two main sources of light (besides the existing fluorescent that was used as control condition, and turned off in the experimental condition.
Figure 4: ED-STAR room after the lighting intervention
Final configuration of the system:
• Two full spectrum light engines in the front side of the room shining
upwards at above average eye level to prevent direct eye exposure
to the LED light sources
• A perimeter lighting profile to provide indirect lighting and uniformity
perception across the room
Both sources of light had different control systems and were commanded through the same communication and control bus (through standard DMX512 protocol). A schematic of the system is depicted in Figure 5 where we can see the two lighting sources and their associated control systems, the DMX communication bus, system controller (lighting controller), the networking interface and the control software interfaces (android/ iPhone “apps” and PC software). The reason to use the wireless network layer instead of a cabled layer for the control was to provide full electrical isolation between the user interface and the lighting system (lighting as a medical device safety).
Figure 5: Lighting system schematic
Light quality
The light quality of the two lighting conditions is shown in figures 6 and 7. The measurements were taken with a calibrated Konica- Minolta CL-500A spectrophotometer. We can see that even if the experimental condition was with a high “blue” content, and at the limit of the black body line (BBL) to be considered “white” (78 000 K), the color fidelity was higher than the conventional fluorescents. Of particular interest for the health care environment is the R9 parameter associated with “red” perception (skin/blood). We did not calculate the “Cyanosis Observation Index” (COI) from Australian regulations (AS/NZS 1680.2.5), as it does not affect USA and EU countries.
Figure 6: Control condition light quality
Figure 7: Experimental condition light quality
Lighting System Commissioning in the Clinical Health care Environment
Once the system was built and installed, we proceeded with the commissioning (programming). Our target for the experiment was to have the same illuminance levels (lux) at the task under both lighting conditions (880 lx), as we can see in figure 8.
Figure 8: Equilux lighting conditions control (left) and experimental (right)
The fluorescent lighting had no control, besides the classical on/off, and we manually adjusted the new experimental lighting system until it achieved the desired experimental condition, with the constraint of providing the same illuminance as the fluorescent system (at the task). This setting was saved as a preset in the lighting system controller (Nicolaudie Stick KE1 with network adapter). The ED-STAR staff was able to setup the room with the push of a button (“mode 5”) in the smartphone “app”. This was demonstrated to be a very robust procedure with no possible mistakes. The staff was very confident about this way of operation.
Subjects: Study population
The research study enrolled emergency department clinicians following the Institutional Review Board (IRB) approved protocol. The study was performed in the Emergency Department Simulation Teaching and Research Center (EDSTAR), a teaching simulation center at Mount Sinai Hospital in NYC, NY, USA. Our experimental setup is a permanent installation that will enable further research with the goal of rapid translatability into the real ED clinical environment.
Statistical model: AB/BA crossover
An AB/BA crossover study is defined as one experiment or clinical trial in which there is an interest in determining differences between two treatments that are given in alternate order to two groups of participants. One group, AB, received first treatment A and then treatment B. The other group, BA, received the treatments in reverse sequence (first B and then A) [73,74,57,51]. Each participant on a crossover acts as his/her own control eliminating the between-participant’s variability, needing fewer observations than the parallel design to get the same precision in estimation. A crossover study is therefore more efficient than a parallel group study.
The statistical model for the AB/BA crossover is summarized in figures 9 and 10 [73]. Figure 9 shows the four cells, with the means and expectations for each cell.
Figure 9: Cell means and expectations for a cross over design
The variables in this figure are:
• General level parameter: µ
• Treatment contrast (treatment effect): τ = τA – τB
• Period effects: π1 and π2
Figure 10, shows the estimation of the parameters of interest (“treatment effect” and “carry-over”) based on the four cell means Y11, Y12, Y21, Y22 (where Y11 is the mean of the first group/sequence, AB, in period 1, and so on).
Figure 10: Estimation of parameters of interest from the cell means for a cross over design
Continuous and discrete variables:
In a crossover study, we have to differentiate carefully between continuous and discrete variables.
Discrete variables:
For the discrete variables, we did a study based on contingency tables.
Continuous variables:
For the continuous variables, we evaluated first the treatment effect (“effect size”) provided by the crossover analysis and the 95% confidence interval. The “effect size” provides the magnitude of the treatment effect and the sign indicates the direction of the effect (higher or lower under experimental compared to control). Then we proceed with the null hypothesis statistically significance testing: Null hypothesis H0: mean difference is zero (no statistical significant difference) Alternate hypothesis H1: we have two different alternate hypothesis.
The two different alternate hypotheses are:
• Mean difference is less than zero (one sided) for:
° KSS (“Karolinska Sleepiness Scale”)
° NASA-TLX (“Task Load Index”)
° Clinical procedures “execution time”
• Mean difference is greater than zero (one sided) for:
° Clinical procedures “performance”
The reason for the two different criteria for the alternate hypothesis is due to the expected hypothesized effects:
• For KSS (sleepiness/alertness) we expect a lower sleepiness
(higher alertness) with the experimental treatment. Same with NASA-TXL
where a lower score will indicate a lower perceived workload. Lower
“execution time” (in seconds) for the clinical procedures is also expected
under the experimental condition
• For clinical procedures “performance” evaluation we hypothesize to have
higher scores under experimental conditions compared to control
We calculate the “p-value” that will be communicated as exact value [79,36,85]. With the treatment effects information (magnitude and direction) and the statistical significance, we will discuss the results and relevance in the next section.
Example of Communication of Crossover Results for Continuous Outcomes:
As there is no standardized way for communicating crossover study results, besides the tables, we will summarize the results in the following format.
In the crossover analysis of outcome yyy, measured zzz, (n = ww) we obtained the following results:
• Treatment effect = x.xx95% CI [x.xx, x.xx]
• Mean C (Control) was higher/ lower (MC = x.x, SDC = x.xx) than
Mean E (Experimental) (ME = x.x, SDE = x.xx), by (xx%),
t(DF) = x.xx, p = .xxx
Experimental Runs, Results, and Data Analysis
Experimental conditions
All participants had a minimum “wash-out” period of at least seven (7) days between experimental runs in order to preclude “carry-over” effects. “Carry-over” can be associated with the learning process (cognitive/psychological) more than the physiological “carry-over” of typical drug crossover studies. Participants were not on night-shifts for at least two days prior to either of the two treatments. The ideal situation would have been to avoid all night-shifts, but this could not be accomplished with active ED clinicians, and it is also closer to the real conditions of the ED environments where experimental results will be translated. We also arranged the timing of the experiment in terms of days of the week and time frames within the day (noon-4pm). All the experiments were conducted during weekdays, excepting Mondays to avoid the potential “post-weekend” effect, also known as “Monday effect” [10].
Sample/participants
Our sample was composed of ten active clinicians (n = from the ED department at Mount Sinai Hospital in NYC (even if ten subjects could be seen initially as a low number, the fact that they were actual clinicians in a realistic clinical environment adds robustness to our study, that is exploratory). All of the participants were volunteers and did not receive any compensation for participation. They were very helpful in accommodating their busy and sometimes unpredictable working schedule to the experimental research agenda, with the strict limitation of the two night shifts.
Characterization of the sample:
• Size: ten (n = 10) participants (equivalent to 20 on a double
arm experiment)
• Experimental runs: twenty (20), two per participant, following the
AB/BA crossover design
• Age: average 35.10 years (SD = 4.95, min = 28, max = 44)
• Gender: seven males and three females
• Experience level: five residents (3th and 4th year) and five
faculty members
• All participants were healthy, and none declared having diabetic
conditions, nor color vision problems
• One participant had light-colored eyes: There was no participant drop-out
Experimental results and data analysis
The data collection and initial processing of the data was done in “MS Excel” and the statistical analysis was performed using “Minitab v17”. As the experimental design was AB/BA crossover, we have used the crossover statistical analysis methodology for the inferential statistics analysis of the data. The crossover design has the unique characteristic of having each participant as its own control, reducing the variability between participants and the sample size required compared to double-arm designs. Besides the conventional inferential study of “p-values”, statistical significance, that has become controversial, yet pervasive, we have also conducted a study of the “treatments effects” to discuss practical significance [56,21,40,55,79,77,22,61,36,85].
Experimental data for KSS and NASA-TLX is shown in different figures and tables in this section: KSS differences within test in Table 1 and Figure 11, the KSS equivalence test in Figure 14; NASA-TLX differences within test in Table 2 and Figure 12, the NASA-TLX equivalence test in Figure 15; crossover data analysis results summary for KSS and NASATLX in Figure 13; the summary of inferential statistics in Table 3. Detailed data for clinical performance execution time and evaluation and for the surveys is not presented here.
Table 1: Within participants’ experimental runs KSS differences within test
Figure 11: Within participants’ experimental runs KSS differences within test (1 = baseline (pre-test), 2 = after procedures (post-test)): Control (left) vs. Experimental (right). Dotted lines represent more than one test with exact results: two 3-3 and two 6-6 occurrences under control (left); two 7-4, two 5-4 and two 4-4 under experimental (right)
Table 2: Within participants’ experimental runs NASA_TLX differences within test (missing data for participant #2 experimental condition)
Figure 12: NASA-TLX within participants (Control vs. Experimental: TLX-C vs.
TLX-E) (missing data for participant #02 experimental condition)
Figures 13: Crossover data analysis results summary for KSS and NASA-TLX
Figures 14: KSS Equivalence Test from Minitab
Figure 15: NASA-TLX Equivalence Test from Minitab (missing data for participant #02 experimental conditions)
The observations of reduction of sleepiness (KSS) and workload (NASA-TLX), factors that may be associated with clinician wellness/ stress and medical error causation in the health care environment, are of special interest in the emergency setting. In the ED environment, clinicians work under high levels of stress [19]. The potential increased ability of saving lives is a major component of the practical significance of our study.
The experimental lighting environment was positively accepted by the study participants and our indirect, “blue-enriched”, full visible spectrum, solid state, “white” lighting (experimental condition) was evaluated in our sample as better than the existing and prevalent fluorescent lighting (control condition). There were slight differences in perception of temperature and humidity: the environment was perceived as more humid and warmer under control (fluorescent) conditions. The experimental condition was accepted by all the participants of our study, except one, and had less claims for controllability. Also, the experimental condition was evaluated as better than the control in the “compared to other workplaces” survey. It is the first time in lighting research that “blue-enriched” full visible daylight lighting delivery has been reported under experimentation with an equivalent CCT of ~ 78 000 K (seventy-eight thousand Kelvin) at the task. This setting was considered better than the conventional lighting (control condition).
We observed in our sample a reduction in sleepiness (alertness? increase) in the KSS scale (-15.94% observed effect with p-value = 0.022), Figure 11, and a reduction in workload perception in the NASA-TLX sum of scales (-21.87% observed effect with p-value = 0.009), Figure 12. We cannot claim a statistically significant reduction in execution times for procedure two and total execution times (-21.04% observed effect with p-value = 0.094). Also, there was no improvement seen in the clinical performance evaluation, even if we can see an observed effect improvement in our study in the scores from control to experimental lighting conditions. Finally, two medical errors occurred under control conditions and none under experimental conditions.
Table 3: Summary of inferential statistics
Discussion
We have obtained results from our experimental research that are supportive of our hypothesis about potential “non-visual”/”non-imageforming” (NIF) beneficial effects of “blue-enriched” full visible spectrum “white” lighting in clinician wellness and performance, and improvement in patient safety. The observed reduction in sleepiness (increase in alertness? wakefulness?) and the observed reduction in perceived workload, the effect in clinical performance, and the acceptance and preference of the experimental condition (indirect, “blue-enriched”, full visible spectrum, solid state, “white” lighting) compared to the control condition (fluorescent lighting) are discussed here. These results may contribute to the translatability of the findings from the simulated to the actual clinical environment.
Our experimental lighting condition suggested an aggregate set of beneficial results for clinician wellness and performance that supports our research hypothesis. Using the existing fluorescent lighting as control reinforced the validity of our study, as these are ubiquitous conditions in the current health care environment.
The acceptance of the experimental lighting condition is critical for its practical feasibility. Also, our lighting conditions provided general lighting performance, while results from previous research that has been conducted in lightboxes [71] or on “ad-hoc” laboratory setups (static subjects with fixed chin) [8] are difficult to be translated into real-world environments. The observed reduction in sleepiness, associated with an increase in alertness (wakefulness?), might initially seem inconsistent with the observed reduction in workload perception, as alertness could be erroneously associated with stress and arousal. In our study, the participants interviewed agreed that the experimental lighting provided a “calming effect” that empowered them to better focus on the execution of the clinical procedures. This “calming effect”, which at the same time improves alertness (wakefulness?) and focus, requires further research and analysis (explored in fields such as mindfulness). Another potential explanation can be found in the (controversial?) field of “Syntonic Optometry” where the effect of “blue” visible radiation is associated with the activation of the parasympathetic nervous system activation (PSNS), [76,52,33]. The field of pharmacodynamics [28] becomes relevant here as perhaps the same wavelength/relative SPD can create different effects depending on levels/dose (and timing).
Taking into account that we performed all of our experiments between noon and 4pm (trying to have a similar circadian time, “CT”, lapse for the participants), we presume that melatonin suppression, or reduction, was not the cause of the effect measured. Melatonin levels are high in the human body during the night and remain low during the day [43]. In the selection of our timeframe we were also careful to not target the morning peak of cortisol. Considering that our timeframe was coincident with the “post-prandial” dip, previous research suggests that the orexin/hypocretin neurons that are associated with wakefulness [25,15,26,72,20] might be inhibited by higher glucose levels associated with high carbohydrate meals [9]. Monk [58] recognizes the post-meal effect, and also the carbohydrates factor, but also proposes that the decrease in performance in the early afternoon could be driven by a 12-hour circadian harmonic that could also be higher in morningtype individuals, and that is not necessarily caused by the post-meal event. Monk also proposes (“bright”) light as a coping mechanism for the “siesta” effect. To the best of our knowledge, only one previous lighting research [66] considers the orexin wakefulness mechanisms in association with lighting, beyond the effects of light in melatonin.
This is clearly an area of future research related to the study of the neurological mechanism underlying the effects of light in human psycho-physiology that would be difficult to conduct in other organisms (such as mice) due to the high level cognitive functions associated with the research, and the fact that mice are nocturnal creatures. Techniques such as fMRI (functional Magnetic Resonance Imaging) have been proposed in the field of neuroscience, even if the statistical methods in neuroimaging have been put recently into question [31]. Other neuro-physiological metrics, such as heart rate variability (HRV), eye tracking (pupillary response), electroencephalography (EEG), and skin conductivity, look very promising to evaluate sympathetic and parasympathetic nervous system activation and should be explored in future studies.
Conclusion and Future Work
We have seen in this exploratory research that controlled lighting environment can be beneficial in the health care environment, particularly in the execution of ED clinical procedures. These environments can improve clinician wellness and performance, and advance patient safety.
This can have a huge impact in the field of health care ergonomics and particularly in the emergent burden of clinician burnout, a topic that has become a priority to the National Academy of Medicine (NAM) [59].
Future work is to expand the scope of our study to advance patient safety in related clinical situations, such as: clinician cognitive recovery from medical error, hand-offs, and improvement of teamwork conditions. These are known clinical scenarios where prevalence of adverse events has been observed, that might be prevented and precluded through environmental interventions such as controlled lighting. Dynamic lighting, temporal effects, ethics, human variability factors, and the interoperability between lighting systems and human objective psycho-physiological variables will be considered.
Acknowledgements:
This research was partially supported by the National Institutes of Health (NIH), National Center for Advancing Translational Sciences (NIH/NCATS), through the 4D Technology Development program at Mount Sinai Hospital, and awarded for its successful completion at the Hospital. The lighting design was shortlisted for the 2016 darc-awards in London, UK.
We thank Griendy Indig, ED-STAR, Mount Sinai Hospital, Icahn School of Medicine, for subject recruitment, lab setup, coordination; Nicholas Genes, MD, emergency medicine clinician for support with Mount Sinai Innovations Partners (MSIP); Rensselaer Polytechnic Institute (RPI) Dean of Graduate Education, Dr. Stanley Dunn, as the director of the doctoral research of Dr. Octavio L. Perez.
We would also like to show our gratitude to Professors Justiniano Aporta and Ana Sanchez-Cano, from the School of Science, Applied Optics, of the University of Zaragoza, Spain, for reviewing the document.
Results of the presented work have been published in the doctoral dissertation of Dr. Octavio L. Perez (under publication embargo until December 2018).
Citations:
[l] “To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able
to say what the experiment died of.” Ronald Fisher (1890-1962)
References:
[1] AHRQ. 2008. “Healthcare Expenses for Adults with Chronic Conditions,
2005.” Statistical Brief #203. Rockville, MD: Agency for Healthcare
Research and Quality
[2] Akerstedt, T, and Mats Gillberg. 1990. “Subjective and Objective
Sleepiness in the Active Individual.” International Journal of
Neuroscience 52 (1-2): 29-37
[3] Anjali, Joseph. 2006. “The Impact of Light on Outcomes in Healthcare
Settings.” Issue Paper #2. Concord, CA: The Center for Health Design
[4] Beauchemain, K. M., and Peter Hays. 1996. “Sunny Hospital Rooms
Expedite Recovery from Severe and Refractory Depressions.” Journal
of Affective Disorders 40 (1-2): 49.
[5] Beauchemin, Kathleen M., and Peter Hays. 1998. “Dying in the Dark:
Sunshine, Gender and Outcomes in Myocardial Infarction.” Journal of
the Royal Society of Medicine 91 (7): 352-354
[6] Benedetti, Francesco, Cristina Colombo, Barbara Barbini,
Euridice Campori, and Enrico Smeraldi. 2001. “Morning Sunlight
Reduces Length of Hospitalization in Bipolar Depression.” Journal of
Affective Disorders 62 (3): 221-223
[7] Berson, David M., Felice A. Dunn, and Motoharu Takao. 2002.
“Phototransduction by Retinal Ganglion Cells That Set the Circadian
Clock.” Science 295 (5557): 1070-1073
[8] Brainard, George C., John P. Hanifin, Jeffrey M. Greeson,
Brenda Byrne, Gena Glickman, Edward Gerner, and Mark D. Rollag.
2001. “Action Spectrum for Melatonin Regulation in Humans: Evidence
for a Novel Circadian Photoreceptor.” The Journal of Neuroscience 21
(16): 6405-6412
[9] Burdakov, Denis, Lise T. Jensen, Haris Alexopoulos, Rhiannan H.
Williams, Ian M. Fearon, Ita O’Kelly, Oleg Gerasimenko, Lars Fugger,
and Alexei Verkhratsky. 2006. “Tandem-Pore K+ Channels Mediate
Inhibition of Orexin Neurons by Glucose.” Neuron 50 (5): 711-22
[10] Butler, Richard J., Nathan Kleinman, and Harold H. Gardner. 2014. “I
Don’t like Mondays: Explaining Monday Work Injury Claims.” Industrial
& Labor Relations Review 67 (3 suppl): 762-783
[11] Canellas, Francesca, Luisa Mestre, Miriam Belber, Guillem Frontera,
Mary Angeles Rey, and Ruben Rial. 2015. “Increased Daylight
Availability Reduces Length of Hospitalisation in Depressive Patients.”
European Archives of Psychiatry and Clinical Neuroscience 266 (3): 277-80
[12] Carayon, Pascale. 2012. “Emerging Role of Human Factors and
Ergonomics in Healthcare Delivery-A New Field of Application and
Influence for the IEA.” Work 41 (Supplement 1): 5037-5040
[13] Carayon, Pascale, Anping Xie, and Sarah Kianfar. 2013. “Human
Factors and Ergonomics.” In Making Health Care Safer II: An Updated
Critical Analysis of the Evidence for Patient Safety Practices. Evidence
Reports/Technology Assessments 211. Rockville, MD: Agency for
Healthcare Research and Quality
[14] Carayon, Pascale, Tosha B. Wetterneck, A. Joy Rivera-Rodriguez,
Ann Schoofs Hundt, Peter Hoonakker, Richard Holden, and Ayse P.
Gurses. 2014. “Human Factors Systems Approach to Healthcare
Quality and Patient Safety.” Applied Ergonomics 45 (1): 14-25
[15] Carter, Matthew E., Luis de Lecea, and Antoine Adamantidis. 2013.
“Functional Wiring of Hypocretin and LC-NE Neurons: Implications
for Arousal.” Frontiers in Behavioral Neuroscience 7: 1-7
[16] Chellappa, Sarah Laxhmi, Roland Steiner, Peter Blattner,
Peter Oelhafen, Thomas Gtz, and Christian Cajochen. 2011.
“Non-Visual Effects of Light on Melatonin, Alertness and Cognitive
Performance: Can Blue-Enriched Light Keep Us Alert?” Edited by
Michael H. Herzog. PLoS ONE 6 (1): e16429
[17] Choi, Joon-Ho, Liliana O. Beltran, and Hway-Suh Kim. 2012.
“Impacts of Indoor Daylight Environments on Patient Average Length of
Stay (ALOS) in a Healthcare Facility.” Building and Environment
50 (April): 65-75
[18] Choi, Kyungah, and Hyeon-Jeong Suk. 2016. “Dynamic Lighting
System for the Learning Environment: Performance of Elementary
Students.” Optics Express 24 (10): A907
[19] Cheng, Yi Han, Gregory D Roach, and Renee MA Petrilli. 2014.
“Current and Future Directions in Clinical Fatigue Management: An
Update for Emergency Medicine Practitioners: Clinical Fatigue
Management Strategies.” Emergency Medicine Australasia 26 (6): 640-44
[20] Chieffi, Sergio, Marco Carotenuto, Vincenzo Monda, Anna Valenzano,
Ines Villano, Francesco Precenzano, Domenico Tafuri, et al. 2017.
“Orexin System: The Key for a Healthy Life.” Frontiers in Physiology 8 (May)
[21] Coe, Robert. 2002. “It’s the Effect Size, Stupid: What Effect Size Is and
Why It Is Important.” In British Educational Research Association
Annual Conference. Exeter, England
[22] Cumming, Geoff. 2013. Understanding the New Statistics: Effect Sizes,
Confidence Intervals, and Meta-Analysis. Multivariate Applications
Series. New York, NY, United States of America: Routledge
[23] Dai, Jiapei, Jan Van Der Vliet, Dick F. Swaab, Ruud M. Buijs, and
others. 1998. “Human Retinohypothalamic Tract as Revealed by in
Vitro Postmortem Tracing.” The Journal of Comparative Neurology 397
(3): 357-370
[24] Daneault, V., M. Dumont, . Mass, G. Vandewalle, and J. Carrier. 2016.
“Light-Sensitive Brain Pathways and Aging.” Journal of Physiological
Anthropology 35 (1): 12
[25] de Lecea, Luis, Matthew E. Carter, and Antoine Adamantidis. 2012.
“Shining Light on Wakefulness and Arousal.” Biological Psychiatry 71
(12): 1046-52
[26] de Lecea, Luis, and Ramn Huerta. 2014. “Hypocretin (Orexin)
Regulation of Sleep-to-Wake Transitions.” Frontiers in Pharmacology 5
(16): 1-7
[27] DiLaura, David L. 2005. “Light’s Measure: A History of Industrial
Photometry to 1909.” Leukos 1 (3): 75-149.
[28] DiPiro, Joseph T., William J. Spruill, William E. Wade, Robert A. Blouin,
and Jane M. Pruemer. 2010. Concepts in Clinical Pharmacokinetics.
Fifth Edition. Bethesda, MD: American Society of Health System
Pharmacists, Inc
[29] US Department of Energy. 2013. “Implementing Spectrally Enhanced
Lighting.” DOE/EE-1048. Building Technologies Office. US Department
of Energy
[30] Eklund, Neil H., and Peter R. Boyce. 1996. “The Development of a
Reliable, Valid, and Simple Office Lighting Survey.” Journal of the
Illuminating Engineering Society 25 (2): 25-40
[31] Eklund, Anders, Thomas E. Nichols, and Hans Knutsson. 2016.
“Cluster Failure: Why FMRI Inferences for Spatial Extent Have Inflated
False-Positive Rates.” Proceedings of the National Academy of
Sciences 113 (28): 7900-7905
[32] Gardner, Aimee K., Dmitry Nepomnayshy, Craig Reickert,
Denise W. Gee, Ryan Brydges, James R. Korndorffer, Daniel J. Scott,
and Ajit K. Sachdeva. 2016. “The Value Proposition of Simulation.”
Surgery 160 (3): 546-51
[33] Gottlieb, Ray, and Larry Wallace. 2001. “Syntonic Phototherapy.”
Journal of Behavioral Optometry 12/2001 (2): 31-38
[34] Gray, Whitney Austin, Karen S. Kesten, Stephen Hurst,
and Laura Anderko. 2012. “Using Clinical Simulation Centers to Test
Design Interventions: A Pilot Study of Lighting and Color Modifications.”
HERD: Health Environments Research & Design Journal 5 (3): 46-65
[35] Diepen, Hester C. van, Russell G. Foster, and Johanna H. Meijer. 2015.
“A Colourful Clock.” PLoS Biology 13 (5): e1002160
[36] Greenland, Sander, Stephen J. Senn, Kenneth J. Rothman,
John B. Carlin, Charles Poole, Steven N. Goodman,
and Douglas G. Altman. 2016. “Statistical Tests, P Values, Confidence
Intervals, and Power: A Guide to Misinterpretations.” European Journal
of Epidemiology 31 (4): 337-50
[37] Hart, Sandra G., and Lowell E. Staveland. 1988. “Development of
NASA-TLX Task Load Index – Results of Empirical and Theoretical
Research.” In Human Mental Workload. North Holland Press
[38] Hart, Sandra G. 2006. “NASA-Task Load Index (NASA-TLX);
20 Years Later.” In Proceedings of the Human Factors and Engineering
Society (HFES) 50th Annual Meeting, 50:904-908. San Francisco,
California: Sage Publications
[39] Hattar, S., Hsi-Wen Liao, Motoharu Takao, David M. Berson,
and King-Wai Yau. 2002. “Melanopsin-Containing Retinal Ganglion
Cells: Architecture, Projections, and Intrinsic Photosensitivity.”
Science 295 (5557): 1065-70
[40] Hedges, Larry V. 2008. “What Are Effect Sizes and Why Do We Need
Them?” Child Development Perspectives 2 (3): 167-171
[41] Hommes, Vanja, and Marina C. Gimnez. 2015. “A Revision of Existing
Karolinska Sleepiness Scale Responses to Light: A Melanopic
Perspective.” Chronobiology International 32 (6): 750-56
[42] Iyendo, Timothy Onosahwo, Patrick Chukwuemeke Uwajeh,
and Ezennia Stephen Ikenna. 2016. “The Therapeutic Impacts of
Environmental Design Interventions on Wellness in Clinical Settings:
A Narrative Review.” Complementary Therapies in Clinical Practice 24
(August): 174-88
[43] Jennum, Poul, Erik Lykke Mortensen, Steen Gammeltoft,
Katja Linda Waller, Martin Lauritzen, Merete Osler, Birgitte Fagerlund,
and Kirsten Avlund. 2016. “Melatonin and Cortisol Profiles in Late
Midlife and Their Association with Age-Related Changes in Cognition.”
Nature and Science of Sleep 8 (January): 47-53
[44] Joarder, A., A. Price, and M. Mourshed. 2009. “Systematic Study of the
Therapeutic Impact of Daylight Associated with Clinical Recovery.”
In Improving Healthcare Infrastructures through Innovation, 25-31.
Brighton, UK: HaCIRIC
[45] Joarder, A., and A. Price. 2013. “Impact of Daylight Illumination on
Reducing Patient Length of Stay in Hospital after Coronary Artery
Bypass Graft Surgery.” Lighting Research and Technology 45 (4): 435-49
[46] Kaida, Kosuke, Masaya Takahashi, Torbjrn kerstedt, Akinori Nakata,
Yasumasa Otsuka, Takashi Haratani, and Kenji Fukasawa. 2006.
“Validation of the Karolinska Sleepiness Scale against Performance
and EEG Variables.” Clinical Neurophysiology 117 (7): 1574-81
[47] Keis, Oliver, Hannah Helbig, Judith Streb, and Katrin Hille. 2014.
“Influence of Blue-Enriched Classroom Lighting on Students’ Cognitive
Performance.” Trends in Neuroscience and Education 3-4: 86-92
[48] Kohn, Linda T., Janet M. Corrigan, and Molla S. Donaldson. 2000.
To Err Is Human: Building a Safer Health System. Washington, D.C.:
National Academies Press
[49] Levin, Scott, Daniel J. France, Robin Hemphill, Ian Jones,
Kong Y. Chen, Dorsey Rickard, Renee Makowski, and Dominik
Aronsky. 2006. “Tracking Workload in the Emergency Department.”
Human Factors: The Journal of the Human Factors and Ergonomics
Society 48 (3): 526-539
[50] Lewy AJ, Wehr TA, Goodwin FK, Newsome DA, Markey SP. 1980.
“Light Suppresses Melatonin Secretion in Humans.” Science 210
(4475): 1267-69
[51] Li, Tianjing, Tsung Yu, Barbara S. Hawkins, and Kay Dickersin. 2015.
“Design, Analysis, and Reporting of Crossover Trials for Inclusion in a
Meta-Analysis.” Edited by Lamberto Manzoli. PLoS ONE 10 (8):
e0133023. https://doi.org/10.1371/journal.pone.0133023
[52] Liberman, Jacob. 1986. “The Effect of Syntonic (Colored Light)
Stimulation on Certain Visual and Cognitive Functions.” Journal of
Optometric Vision Development 27 (June): 4-14
[53] Lucas, Robert J., Stuart N. Peirson, David M. Berson,
Timothy M. Brown, Howard M. Cooper, Charles A. Czeisler,
Mariana G. Figueiro, et al. 2014. “Measuring and Using Light in the
Melanopsin Age.” Trends in Neurosciences 37 (1): 1-9
[54] Mayron, Lewis W., John Ott, Ricks Nations, and Ellen L. Mayron. 1974.
“Light, Radiation, and Academic Behavior.” Intervention in School and
Clinic 10 (1): 33-47
[55] McGough, James J., and Stephen V. Faraone. 2009. “Estimating the
Size of Treatment Effects: Moving beyond p Values.” Psychiatry
(Edgemont) 6 (10): 21-29
[56] McGraw, Kenneth O., and S. P. Wong. 1992. “A Common Language
Effect Size Statistic.” Psychological Bulletin 111 (2): 361-65
[57] Mills, Edward J, An-Wen Chan, Ping Wu, Andy Vail, Gordon H Guyatt,
and Douglas G Altman. 2009. “Design, Analysis, and Presentation of
Crossover Trials.” Trials 10 (27): 1-6
[58] Monk, Timothy H. 2005. “The Post-Lunch Dip in Performance.” Clinics
in Sports Medicine 24 (2): e15-23
[59] National Academy of Medicine. 2017. “Action Collaborative on Clinician
Well-Being and Resilience.”
https://nam.edu/initiatives/ clinician-resilience-and-well-being/
(last accessed 2018-06-01)
[60] NASA. 2015. “NASA Technology Roadmaps – TA6: Human Health,
Life Support, and Habitation Systems.” 2015 TA 6. NASA
[61] Nuzzo, Regina, and others. 2014. “Statistical Errors.” Nature 506
(7487): 150-152
[62] O’Hagan, J B, M Khazova, and L L A Price. 2016. “Low-Energy Light
Bulbs, Computers, Tablets and the Blue Light Hazard.” Eye 30 (2): 230-33
[63] Provencio, Ignacio, Guisen Jiang, Willem J. De Grip, William Par
Hayes, and Mark D. Rollag. 1998. “Melanopsin: An Opsin in
Melanophores, Brain, and Eye.” Proceedings of the National Academy
of Sciences, Neurobiology, 95 (1): 340-45
[64] Provencio, Ignacio, Ignacio R. Rodriguez, Guisen Jiang, William Pr
Hayes, Ernesto F. Moreira, and Mark D. Rollag. 2000. “A Novel Human
Opsin in the Inner Retina.” The Journal of Neuroscience 20 (2): 600-605
[65] Provencio, Ignacio. 2011. “The Hidden Organ in Our Eyes.”
Scientific American 304 (5): 54-59
[66] Rautkyl, Emilia. 2011. “Research Methods and Neurophysiological
Mechanisms behind the Alerting Effects of Daytime Light Exposure.”
Doctoral Dissertation, Helsinki, Finland: Aalto University
[67] Rahman, Shadab A., Melissa A. St Hilaire, Claude Gronfier,
AnneMarie Chang, Nayantara Santhi, Charles A. Czeisler,
Elizabeth B. Klerman, and Steven W. Lockley. 2018.
“Functional Decoupling of Melatonin Suppression and Circadian Phase
Resetting in Humans: Decoupled Melatonin Suppression and Phase
Resetting.” The Journal of Physiology 596 (11): 2147-57
[68] Reiling, John, Hughes, Ronda G., and Mike R. Murphy. 2008.
“Chapter 28. The Impact of Facility Design on Patient Safety.” In
Patient Safety and Quality: An Evidence-Based Handbook for Nurses.
Rockville, MD: Agency for Healthcare Research and Quality
[69] RCGP. 2015. “Patient Safety Implications of General Practice
Workload.” London, England: Royal College of General Practitioners
[70] Sadun, Alfredo A., Judith D. Schaechter, and Loise E.H. Smith. 1984.
“A Retinohypothalamic Pathway in Man: Light Mediation of Circadian
Rhythms.” Brain Research 302: 371-77
[71] Sahin, Levent, and Mariana G. Figueiro. 2013. “Alerting Effects of
Short-Wavelength (Blue) and Long-Wavelength (Red) Lights in the
Afternoon.” Physiology & Behavior 116-117: 1-7
[72] Schwartz, Michael D., and Thomas S. Kilduff. 2015. “The Neurobiology
of Sleep and Wakefulness.” Psychiatric Clinics of North America 38
(4): 615-44
[73] Senn, Stephen. 1996. “The AB/BA Cross-over: How to Perform the
Two-Stage Analysis If You Can’t Be Persuaded That You Shouldn’t.”
In Liber Amicorum Roel van Strik, 93-100. Rotterdam: Erasmus University
[74] Senn, Stephen. 2002. Cross-over Trials in Clinical Research.
Second Edition. Statistics in Practice. Chichester, England:
John Wiley & Sons, Ltd
[75] Shahid, Azmeh, Kate Wilkinson, Shai Marcu, and Colin M. Shapiro.
2011. “Karolinska Sleepiness Scale (KSS).” In STOP, THAT and One
Hundred Other Sleep Scales, edited by Azmeh Shahid, Kate Wilkinson,
Shai Marcu, and Colin M Shapiro, 209-10. New York, NY: Springer New York
[76] Spitler, H.R. 1941. The Syntonic Principle – Its Relation to Health and
Ocular Problems. Eaton, OH: College of Syntonic Optometry
[77] Sullivan, Gail M., and Richard Feinn. 2012. “Using Effect Sizeor Why
the P Value Is Not Enough.” Journal of Graduate Medical Education 4
(3): 279-82
[78] Thapan, Kavita, Josephine Arendt, and Debra J. Skene. 2001.
“An Action Spectrum for Melatonin Suppression: Evidence for a
Novel Non-Rod, Non-Cone Photoreceptor System in Humans.”
The Journal of Physiology 535 (1): 261-267
[79] Thirthalli, Jagadisha, and Ravi Philip Rajkumar. 2009.
“Statistical versus Clinical Significance in Psychiatric Research An
Overview for Beginners.” Asian Journal of Psychiatry 2 (2): 74-79
[80] Ulrich, Roger. 1984. “View through a Window May Influence Recovery.”
Science 224 (4647): 224-225.
[81] Vandewalle, Gilles, Christina Schmidt, Genevive Albouy,
Virginie Sterpenich, Annabelle Darsaud, Graldine Rauchs,
Pierre-Yves Berken, et al. 2007. “Brain Responses to Violet, Blue,
and Green Monochromatic Light Exposures in Humans:
Prominent Role of Blue Light and the Brainstem.” PLoS ONE 2 (11): e1247
[82] Veitch, Jennifer A., and Shelley L. McColl. 2001. “A Critical
Examination of Perceptual and Cognitive Effects Attributed to
FullSpectrum Fluorescent Lighting.” Ergonomics 44 (3): 255-79
[83] Walch, Jeffrey M., Bruce S. Rabin, Richard Day, Jessica N. Williams,
Krissy Choi, and James D. Kang. 2005. “The Effect of Sunlight on
Postoperative Analgesic Medication Use: A Prospective Study of
Patients Undergoing Spinal Surgery” Psychosomatic Medicine 67
(1): 156-63
[84] Wallace, Jean E., Jane B. Lemaire, and William A. Ghali. 2009.
“Physician Wellness: A Missing Quality Indicator.” The Lancet 374
(9702): 1714-1721
[85] Wasserstein, Ronald L., and Nicole A. Lazar. 2016. “The ASA’s
Statement on p-Values: Context, Process, and Purpose.” The American
Statistician 70 (2): 129-33
[86] Wild, C. P. 2005. “Complementing the Genome with an “Exposome”:
The Outstanding Challenge of Environmental Exposure Measurement
in Molecular Epidemiology.” Cancer Epidemiology Biomarkers &
Prevention 14 (8): 1847-50
[87] Zeitzer, Jamie M. 2018. “When Is a Proxy Not a Proxy? The Foibles of
Studying Non-Image Forming Light: Perspectives.” The Journal of
Physiology 596 (11): 2029-30
[88] Zhao, Mingjie, Yang-Seon Kim, and Jelena Srebric. 2015.
“Occupant Perceptions and a Health Outcome in Retail Stores.”
Building and Environment 93 (November): 385-94
