HLTH 4200 Week 10 Causal Criteria Essay Example

Does Handheld Phone Use Cause Crashes Among Drivers Under 25? Hill's Considerations Sorted Into Those Three Composite Studies Address and Those They Cannot, and Rothman's Causal Pies Explaining Why the Association Varies

Student Name

Walden University

HLTH 4200: Principles of Epidemiology

Instructor Name

Month Day, Year

Does Handheld Phone Use Cause Crashes Among Drivers Under 25? Hill's Considerations Sorted Into Those Three Composite Studies Address and Those They Cannot, and Rothman's Causal Pies Explaining Why the Association Varies

The Association

Across three composite studies summarized below, drivers under 25 who were using a handheld phone were more likely to crash than when they were not. The question is whether that association reflects cause. This essay applies the considerations proposed by Hill (1965), sorting them into those the studies address and those they cannot, and then uses the component-cause model proposed by Rothman (1976) to explain why the association is stronger in some settings than in others. The three studies are invented for this exercise, but each resembles a design used in published research, and the essay notes the real studies they resemble.

The Three Composite Studies

Study 1 is a case-crossover study of 500 drivers aged 18 to 24 who were in police-reported crashes. Each driver's phone records were compared for the 10 minutes before the crash and the same 10 minutes on a trip the previous week. Handheld phone use in the minutes before the crash was associated with an odds ratio of 4.0. A real study of this design, using billing records, found a fourfold risk of collision during phone calls among drivers of all ages (Redelmeier & Tibshirani, 1997), and another, interviewing drivers in hospital, found an odds ratio of 4.1 for phone use up to 10 minutes before an injury crash (McEvoy et al., 2005).

Study 2 is a naturalistic driving study of 120 newly licensed drivers whose cars were fitted with cameras and sensors for 18 months. Dialing a handheld phone was associated with an odds ratio of 6.5 for a crash or near-crash, compared with driving without secondary tasks. A real study using this design found an odds ratio of 8.32 for dialing among novice drivers (Klauer et al., 2014).

Study 3 is an ecological comparison of states. Crash rates among drivers aged 16 to 24 fell by 5% in the two years after five states banned handheld phone use while driving, compared with no change in five similar states without bans.

Considerations the Studies Address

Strength

Hill (1965) argued that strong associations are less likely to be explained entirely by bias or confounding. Odds ratios of 4.0 and 6.5 are strong. A confounder would need to be both very common during phone use and strongly linked to crashes to explain them away.

Consistency

The association appears in three different designs, with different populations, data sources and analytic approaches. The case-crossover study relies on phone records, the naturalistic study on video and the state comparison on police crash counts. Each design has different weaknesses, so agreement among them is more persuasive than agreement among three studies of one type.

Temporality

The cause must come before the effect. Study 1 measures phone use in the minutes before the crash, and Study 2 observes it directly on video before each event. Temporality is established for these two. In Study 3, the bans came before the decline, though other changes may have happened at the same time.

Biological Gradient

A dose-response relationship strengthens a causal reading. Study 2 can address this: tasks requiring more eyes-off-road time, such as dialing and texting, carry higher odds than talking. Study 1 could test whether calls closer to the crash carry more risk; the real case-crossover study found calls within 5 minutes of the collision especially hazardous (Redelmeier & Tibshirani, 1997).

Plausibility

A mechanism is available. Holding and operating a phone takes a hand from the wheel and the eyes from the road, and conversation draws attention. The mechanism is consistent with what is known about attention and reaction time.

Experiment

Hill asked whether removing the exposure changes the outcome. Study 3 approximates this: when bans reduced exposure, crashes among young drivers fell. It is not a randomized experiment, and the decline is modest.

Considerations the Studies Cannot Address

Specificity

Hill noted that a cause that leads to one specific effect strengthens the argument, but he also warned that the absence of specificity does not weaken it much. Phone use is not specific to crashes, and crashes have many causes. The studies cannot satisfy this consideration, and it matters little here.

Coherence and Analogy

Coherence asks whether a causal reading conflicts with what is known about the natural history of the outcome, and analogy asks whether similar exposures have similar effects. Other distractions, such as eating or reaching for objects, are also linked to crashes; the real naturalistic study found elevated odds for several such tasks among novice drivers (Klauer et al., 2014). This supports coherence, but the three composite studies themselves were not designed to test other distractions, so analogy rests on outside evidence.

Confounding Within the Ecological Study

Study 3 cannot show that the drivers who stopped using phones were the drivers who stopped crashing. States that pass bans may also change speed enforcement or graduated licensing at the same time. The ecological design supports the causal argument only weakly on its own.

Rothman's Causal Pies: Why the Association Varies

Rothman (1976) described causes as components of sufficient causes. A sufficient cause is a set of conditions that together produce the outcome; each set can be drawn as a pie, with each component a slice. A crash might be produced by one pie containing handheld phone use, inexperience and a sudden stop by the car ahead; another pie might contain fatigue, a curving road at night and speed. Phone use is a component cause in some pies and absent from others.

This model explains why the strength of the association varies between settings. The odds ratio for phone use depends on how common the other components of its pies are. Among novice drivers, inexperience, one complementary component, is always present, so phone use more often completes a sufficient cause, and the association is stronger, as the higher odds ratio in Study 2 suggests. On a quiet rural road with few sudden stops, the complementary components are rarer, and the same phone use completes fewer pies. The causal pies also explain why bans produced only a modest decline in Study 3: removing one component prevents only the crashes whose pies included it, and many crashes come from pies without phone use.

Conclusion

Hill's considerations of strength, consistency, temporality, gradient, plausibility and experiment are addressed by the three composite studies and point toward a causal role for handheld phone use in crashes among young drivers. Specificity, analogy and the ecological study's confounding cannot be settled by these studies. Rothman's causal pies show that phone use need not cause crashes on its own to be a cause: it is a component that completes some sufficient causes, more often among inexperienced drivers, which is why the association is strongest where they drive.

References

Hill, A. B. (1965). The environment and disease: Association or causation? Proceedings of the Royal Society of Medicine, 58(5), 295-300. https://doi.org/10.1177/003591576505800503

Klauer, S. G., Guo, F., Simons-Morton, B. G., Ouimet, M. C., Lee, S. E., & Dingus, T. A. (2014). Distracted driving and risk of road crashes among novice and experienced drivers. New England Journal of Medicine, 370(1), 54-59. https://doi.org/10.1056/NEJMsa1204142

McEvoy, S. P., Stevenson, M. R., McCartt, A. T., Woodward, M., Haworth, C., Palamara, P., & Cercarelli, R. (2005). Role of mobile phones in motor vehicle crashes resulting in hospital attendance: A case-crossover study. BMJ, 331(7514), 428. https://doi.org/10.1136/bmj.38537.397512.55

Redelmeier, D. A., & Tibshirani, R. J. (1997). Association between cellular-telephone calls and motor vehicle collisions. New England Journal of Medicine, 336(7), 453-458. https://doi.org/10.1056/NEJM199702133360701

Rothman, K. J. (1976). Causes. American Journal of Epidemiology, 104(6), 587-592. https://doi.org/10.1093/oxfordjournals.aje.a112335