Of the 1,788 retired NFL players in the sampling pool, 599 (33.5%) had incorrect or disconnected telephone numbers; five (0.3%) died prior to study commencement. The eligible sample, therefore, contained 1,184 individuals. Of these, 83 (7%) refused and 457 (38.6%) were unreachable. A total of 644 players (54.4%) completed the survey and are the focus of these analyses. Although more players might have been reachable with additional time, we exhausted the limited funds available for the effort. The largest proportion of players was from the southeast (37%) with 21% from the west coast, 19% from the midwest, 14% from the southwest, and 9% from the northeast.
3.1 Sample characteristics
As shown in , the players were 48.3 years of age (SD 9.24) on average; 52% were Caucasian and 45% African American. Offensive lineman was the position most frequently played (28%). On average, players were retired 17.7 years (SD 8.47) and reported playing 7.6 years (SD 3.8). Most (88%) felt they were in excellent health at the start of their careers, but only 13% felt the same in the past 30 days.
Characteristics of Retired NFL Players by Opioid Use Status While Playing for the NFL
Knee injuries were the most commonly reported NFL injuries, followed by shoulder and back. Nearly half (47%) had 3 or more NFL injuries. Concussions from NFL play were prevalent: nearly half (49%) reported diagnosed concussions; 81% reported undiagnosed concussions. The average number of reported concussions of either type was 9 (SD 20.4). Over half (55%) reported a career-ending injury with 6% reporting current use of a cane, walker, or wheelchair. Additionally, players perceived that nearly one-third (29.4%) of their teammates misused prescription opioids. Of the 644 players in the sample, 14 (2.2%) reported that someone told them that they had a problem with opioid medications.
3.2 Trajectory of use and misuse from NFL to present
Shown in is the trajectory of use from the NFL through the interview period; 52% (n=336) of retired NFL players reported using opioids during their NFL careers. Of those players, 37% obtained their opioids exclusively from a doctor, 12% got them exclusively from a non-medical source, and the remaining majority (51%), reported the source of their prescription opioids to be a combination of both doctors and illicit sources such as a teammate, coach, athletic trainer, or family member.
History of Prescription Opioid Use from NFL to Present
Of the players who used prescription opioids in the NFL (n=336), 71% misused in the NFL. Among NFL misusers, 17% (n=40) used just as prescribed in the past 30 days, 68% (n=163) reported no use, and 15% (n=35) reported misuse. Among those who used just as prescribed in the NFL, 8% (n=8) also used just as prescribed in the past 30 days, 87% (n=85) reported no use, and 5% (n=5) reported misuse. Thus, players who misused during their NFL career were 3.2 times as likely to misuse in the past 30 days as NFL players who used just as prescribed (95%CI 1.22-8.44). Among those players who reported no opioid use during their NFL careers (n=308), 93% reported no opioid use in the past 30 days, while 2% and 5% reported misuse and use just as prescribed, respectively. Of the 7% (n=45) of players who reported past 30 day opioid misuse (shaded boxes), 78% (35/45) had a history of opioid misuse during their NFL career.
3.3 Use vs. non-use, during the NFL
Comparisons of opioid use patterns were made between players who used during their NFL career (n=336) and those who did not (n=308). As shown in , younger and Caucasian players had an increased risk for any opioid use during their NFL careers compared to older and non-Caucasian players. Offensive linemen were 2.09 times as likely as all others to use opioids (95%CI 1.47-2.98). Users were more likely to report knee injuries, compared to non-users (67% vs. 54% with OR 1.69, 95%CI 1.23-2.32). Persons with at least 3 NFL injuries and injuries that ended their careers had increased odds for reporting opioid use during their NFL careers compared to persons with a lower injury threshold (OR 2.05, 95%CI 1.50-2.81) or without a career ending injury (OR 1.68, 95%CI 1.23-2.30). While diagnosed concussions did not predict opioid use during players’ NFL careers, undiagnosed concussions did increase the odds for reporting opioid use (OR 2.05, 95%CI 1.35-3.11). Players who used opioids during their NFL careers had been retired fewer years and reported worse current health than those who did not use opioids during their NFL careers. A weak association was found between opioid use during a player's NFL career and the perception that one's teammates misused prescription opioids.
3.4 Misuse vs. use just as prescribed, during the NFL
Though not shown, we compared players who used opioids just as prescribed during their NFL careers (n=98; 29%) to those who misused them (n=238; 71%). Only two differences were found: misusers had increased odds for poor health at retirement (OR .35, 95%CI .16-.73) and had 3 or more NFL injuries (OR 1.85, 95%CI 1.15-2.97).
Additional analyses were performed to further elucidate the association between NFL misuse and other factors. On average, players who misused opioids used them for 5.9 seasons (SD 3.73) compared to 3.7 seasons (SD 2.99) among those who used just as prescribed (p<.0001). While there were no statistically significant differences noted in the number of opioid pills, patches or liquid doses used on a typical day between those who misused and those who used just as prescribed during their careers in the NFL, 22% of misusers reported taking 6 or more pills per day vs. 10% of players who used just as prescribed (p=.01).
3.5 Misuse vs. non-use, in the past 30 days
Comparisons of players by current past 30 day opioid use status () were done for misusers vs. non-users (45 vs. 534), misusers vs. those who used just as prescribed (45 vs. 65), and for users and misusers vs. non-users (110 vs. 534).
Characteristics of Retired NFL Players by Past 30 Day Use of Opioids (N=644)
Misusers were less likely than non-users (C vs. A) to report excellent health in the past 30 days (2% vs. 15%), more likely to report knee, shoulder and back injuries, and over 6 times as likely to report 3 or more NFL injuries (OR 6.02, 95%CI 2.75-13.18). Nearly all (98%) misusers, compared to non-users (79%), reported undiagnosed concussions (OR 11.22, 95%CI 1.53-82.48), the strongest predictor of misuse in the bivariate analysis. Misusers were at increased odds of having a career ending injury and nearly 8 times as likely to be using a cane, walker or wheelchair (OR 7.59, 95%CI 3.16-18.22) compared to their non-using teammates. Nearly 80% of the misusers met criteria for severe pain compared to 39% of non-users (OR 4.50, 95%CI 2.42-8.40). When compared to non-users, misusers reported more self-perceived moderate to severe physical (OR 3.49, 95%CI 1.76-6.91) and mental (OR 2.62, 95%CI 1.38-4.97) impairment.
Players who drank more than 14 drinks in a 7 day period, considered heavy use by the National Institute on Alcohol Abuse and Alcoholism-NIAAA (NIAAA, 2010
), had increased odds of opioid misuse (OR 3.92, 95%CI 1.96-7.81). This association persisted for those drinking 20+ drinks (equal to a fifth of liquor) in the previous week (OR 4.47, 95%CI 2.14-9.33).
Misusers perceived that 43% of their teammates were likely to misuse opioids compared to 25% among non-users.
Although not shown, multivariate regression analysis, predicting current misuse, with non-use as the referent group, indicated three main determinants: moderate to severe pain (AOR 8.33, 95%CI 1.98-35.04), undiagnosed concussions (AOR 3.51, 95%CI 1.98-35.04) and drinking 20+ drinks in the past week (AOR 3.48, 95%CI 1.63-7.41).
3.6 Misuse vs. use just as prescribed, in the past 30 days
A comparison of current misusers to users just as prescribed (C vs. B in ) indicated that players who were younger, retired fewer years, with knee or shoulder injuries, three or more NFL injuries, undiagnosed concussions, and who reported drinking more than 14 drinks per week were at increased odds for past 30 day opioid misuse. However, past 30 day misusers of prescription opioids were no more likely than players who used just as prescribed to be impaired from their injuries and to report more pain.
Multivariate analyses indicated that undiagnosed concussions singly, net of all other variables, predicted opioid misuse (AOR 4.25, 95%CI 1.12-16.22).
3.7 Use vs. non-use, in the past 30 days
When past 30 day users, regardless of misuse patterns, were compared with non-users (BC vs. A in ), many differences were noted, but those with an odds ratio greater than 2.0 included back injury, three or more NFL injuries, undiagnosed concussions, moderate to severe physical and mental impairment and drinking more than 14 drinks per week. However, two additional factors were strongly associated with opioid use: requiring a cane, walker or wheelchair (OR 7.59, 95%CI 3.88-14.86), and having severe pain (OR 4.02, 95%CI 2.62-6.16).
Multivariate analyses increased the strength of the association for moderate to severe pain in predicting opioid use (AOR 6.76, 95%CI 2.88-15.87). Two additional variables remained in the model after backward elimination: moderate to severe mental impairment (AOR 1.88, 95%CI 1.19-2.98) and drinking 20+ drinks in the past week (AOR 2.15, 95%CI 1.17-3.98).