A Corporate Driven Sleep Apnea Detection and Treatment Program: Results and Challenges
Page 2 of 5: Methods
Corporate Measures Facilitating Diagnosis and Treatment
Schneider National Inc (SNI) has recognized SDB with financial backing since 1998. A SNI employed driver referred for SDB testing and physician consultation incurs no out of pocket expenses. Since the late 1990's, the Occupational Health Department at SNI has noticed a trend in their driver population where the risk of SDB and the co-morbidities associated with SDB were evident in 1 out of 3 drivers. This discovery led to an increased effort by the Occupational Health Department to begin a formal process of symptom recognition and SDB education throughout the organization.
Educating a mobile workforce has its own set of challenges. The nurses work to identify every available internal resource to assist them in this regard. Training engineers have become an effective first-line screen of new drivers. From the classroom to riding in the cabs beside these novice drivers, symptoms of sleepiness and diminished vigilance are readily observed. Disease Management vendors handling cardiac disease, diabetes, and asthma are trained to monitor for symptoms of SDB. On site physical therapists at SNI's various operating centers are trained to identify drivers suspected of having SDB. Loss Prevention teams are similarly trained. Text messages are routinely sent to drivers in their trucks, reminding them to schedule a sleep study or a follow-up appointment. A monthly newsletter is sent to the driver's homes featuring articles on SDB (spousal recognition of symptoms is a key diagnostic aid). The medical interview required at the time of hire has modified its strategy to better identify the SDB candidate. Pharmaceutical claims data is reviewed for certain medications, including sleep aids and Modafinil, often triggering a sleep evaluation for those drivers. Successfully treated SDB drivers are interviewed and recorded for educational "drive-time tapes." These educational resources allow SDB-treated drivers to describe to their peers their renewed sense of energy, productivity, and improved sense of safety. Short-term disability waiting periods are waived for those undergoing diagnosis and treatment for SDB. Driver handbooks are updated to include SDB when addressing fatigue over the road.
Despite all of the previously mentioned initiatives, drivers' concern over possible loss of income and job security remained as significant impediments in implementing widespread driver cooperation. Measures were then developed addressing those concerns, including a two-day turnaround from diagnosis to treatment of SDB. A business relationship was formed between Precision Pulmonary Diagnostics, Inc. (PPD) and SNI whereby a sleep diagnosis/treatment facility located near a major SNI operating center was identified to deliver rapid turn-around times from diagnosis to treatment set-up - averaging 2 business days. This center was held to the highest medical and service standards. Using their operating center as a hub, SNI facilitates referral to this sleep facility through preferential route scheduling of drivers in need of testing. PPD also works with SNI in efforts to educate the DOT physicians contracted with SNI to perform driver certification exams. These efforts have included two mass educational mailings to all involved DOT physicians working with SNI, as well as face-to-face meetings with said physicians in the city where a participating SNI operating center is located. Through separate mailings to DOT physicians, SNI has "raised the bar" explicitly identifying what documentation they require of their contracted physicians.
Screening Tool
Certain symptoms carry a near 100% accuracy in predicting a diagnosis of SDB and should prompt a sleep study, such as witnessed apnea by a bed-partner. Other symptoms, when volunteered by the driver or discovered through appropriate survey responses, justify testing for a sleep disorder. One example is daytime sleepiness, with or without a history of snoring. On the other hand, snoring alone may not be sufficient to warrant obtaining the more expensive polysomnography. Trucking companies require a valid and relatively inexpensive tool allowing identification of those individuals with a high likelihood of having SDB and therefore requiring polysomnography. Such an ideal screening tool should be sufficiently sensitive to pick up all truckers with clinically relevant SDB. The screening tool should incorporate relevant subjective response data useful in identifying high risk groups while not completely bound by a respondent's answers to exclude a high risk individual. This latter point is important because some drivers may try to conceal their symptoms of daytime sleepiness out of fear it may jeopardize their employment. Therefore a screening tool must also include reliable, objective data in formulating a prediction for sleep apnea in any given person. From a review of the available medical literature, we developed a simple screening tool to augment the aforementioned subjective responses of the drivers. This tool incorporates weighted values for body mass index (BMI), presence of hypertension, and presence or absence of heavy snoring. The tool also includes other objective criteria for a screener to consider in deciding whether or not to refer a driver for testing.
Accident Rate Analysis
Driver identification numbers in the sample of SDB data was analyzed along side of accident data matching the same driver numbers. A portion of SDB data was not usable due to a dual role of driver and shop mechanic or the original date of hire was not located due to rehiring of that driver. Of the 337 SDB employees analyzed in this report, 255 were deemed acceptable for analysis of accident rates, based on the exclusions mentioned above. Beginning with the driver's date of hire, all preventable accidents prior to treatment for SDB were tracked. A comparison was made looking at the number of drivers with preventable accidents both before and after SDB treatment.
Retention Rate Analysis
This was obtained by noting the percentage of CPAP-treated drivers diagnosed during the study period and still employed by SNI as of June 30, 2005. A comparison was made with the 2004 corporate-wide driver retention rate for SNI (actual numbers were deemed proprietary).
Health Care Analysis
This analysis identified drivers who received a CPAP machine and compared the rates of hospital admissions, emergency room (ER) visits, office visits, and outpatient (OP) visits as well as medical, drug, and total health care spending for these individuals in the periods before and after receiving CPAP intervention. The analysis involved merging claims files from multiple sources: medical claims from Cigna for 2003 and 2004 (provided by Schneider), pharmacy claims for the same time periods (obtained from Medco), and medical and pharmacy claims from Definity Health paid from January through June of 2005. This allowed for the creation of a continuous 30-month claims stream for Schneider from January of 2003 through June of 2005. In the absence of member-level eligibility data, all members were treated as continuously enrolled for the years in which they had claims paid. An alternative to this approach would have been to use the first and last dates for which each member had paid claims, but a check against Definity Health's eligibility data determined that this method underestimated eligibility far more than the assumption of continuous enrollment overestimated it. The study group was identified using a list of DME (Durable Medical Equipment) codes for the CPAP machine and related items among Schneider subscribers. An intervention date was then identified for each study group member by selecting the earliest service date from all claims with one of the appropriate DME codes. This allowed for the calculation of the number of months of enrollment for each member before and after the CPAP intervention. Utilization was identified from both data sources, with different identifiers used due to different fields in the data. Admissions were identified through the presence of an admission date in the Cigna data, and through revenue codes in the Definity Health data. ER visits were found through the use of place-of-service (POS) codes in the Cigna data, and POS codes or revenue codes in the Definity Health data. Office visits were identified using POS codes, and OP visits using minor service category and POS codes. To prepare the dataset for analysis, utilization and spending variables were assigned to a pre or post-intervention row for each member in the study group based upon dates of service. If the CPAP intervention occurred during an admission, the admission and its associated costs were assigned to the pre-intervention period; all other non-admission intervention costs fell into the post period. The Cigna and Definity Health values were then merged and summed, providing a single dataset broken down by pre and post-CPAP values for each member in the study group. In order to accommodate different lengths of enrollment for each member, the number of months of enrollment before and after the intervention were calculated and used as the denominator in PMPM measures of utilization and spending. Finally, members were removed who did not have at least one full month of enrollment before and after the CPAP intervention. The study group contains 348 members. Comparing pre and post-intervention values, summary statistics were gathered for all of the per member per month (PMPM) variables. T-tests were also run to test the statistical significance of the difference between mean pre and post-intervention values (95% confidence interval). Two tests were run, one which assumes equal variances in the pre and post-populations (pooled) and one which does not (Satterthwaite). Both tests produced identical t-values and levels of statistical significance, so only one value is reported in the results for each mean.
CPAP Compliance Data
A subset of CPAP-treated drivers available to PPD for contact was queried regarding their frequency of equipment usage, symptoms of daytime sleepiness, difficulty with CPAP therapy, and various subjective responses related to treatment success.
