There is limited evidence about the prescribing of smoking cessation products to patients recently diagnosed with cancer. In this chapter, we investigate the rate of smoking in patients diagnosed with cancer, the prevalence of prescribing smoking cessation products to patients and the effectiveness of smoking cessation in this population. In the results presented below, we focus on the conventional analysis for two reasons. First, we found little evidence of bias in the effectiveness analysis for the full sample using conventional methods. Second, the sample of patients recently diagnosed with cancer is much smaller than the full sample. Therefore, the statistical power is much lower.
Population characteristics
Of the 220,136 patients prescribed smoking cessation products sampled in the effectiveness analysis in Chapter 3, 2965 had a diagnosis of cancer in the year before their smoking cessation prescription. A smaller proportion of patients diagnosed with cancer (20.9%) were prescribed varenicline than non-cancer patients (32.2%). On average, patients diagnosed with cancer were older than the general population (59 years vs. 46 years), were more likely to be male and had more comorbidities, but were otherwise similar (Table 4).
| Characteristic | Any incident cancer (N = 2965) | No incident cancer (N = 217,171) | ||||
|---|---|---|---|---|---|---|
| NRT (N = 2345) | Varenicline (N = 620) | Total | NRT (N = 147,181) | Varenicline (N = 69,990) | Total | |
| Age (years) at time of first prescription, mean (SD) | 59.6 (12.3) | 56.7 (12.2) | 59.02 (12.3) | 46.2 (15.5) | 44.4 (13.2) | 45.6 (14.8) |
| Sex (male), n (%) | 1195 (51.0) | 313 (50.5) | 1508 (50.9) | 67,983 (46.2) | 34,831 (49.8) | 102,814 (47.3) |
| Multiple deprivation score (IMD), mediana | 12 | 12 | 12 | 12 | 12 | 12 |
| Number of GP visits 1 year prior to first prescription, mean (SD) | 11.8 (9.0) | 11.1 (8.6) | 11.7 (8.8) | 7.8 (7.3) | 6.2 (6.0) | 7.3 (7.0) |
| BMI (kg/m2), mean (SD)a | 25.8 (5.3) | 26.2 (5.5) | 25.9 (5.4) | 26.5 (5.7) | 26.5 (5.4) | 26.5 (5.6) |
| Year of first prescription, median | 2009 | 2010 | 2010 | 2009 | 2010 | 2009 |
| Days of history, mean (SD) | 3662.8 (1996.2) | 3323.5 (2012.4) | 3591.8 (2004) | 3150.7 (1889.3) | 3283.5 (1976.3) | 3193.5 (1918.8) |
| Comorbidity ever (Charlson Comorbidity Index), n (%) | 2050 (87.4) | 492 (79.4) | 2542 (85.7) | 54,224 (36.8) | 22,031 (31.5) | 76,255 (35.1) |
| Alcohol misuse ever, n (%) | 1318 (13.6) | 66 (10.7) | 384 (13.0) | 12,678 (8.6) | 4354 (6.2) | 17,032 (7.8) |
| Drug misuse ever, n (%) | 49 (2.1) | 13 (2.1) | 62 (2.1) | 4546 (3.1) | 1344 (1.9) | 5890 (2.7) |
| Self-harm ever, n (%) | 218 (9.3) | 52 (8.4) | 270 (9.1) | 15,719 (10.7) | 6128 (8.8) | 21,847 (10.1) |
| Ever anxiety and stress-related disorders, n (%) | 663 (28.3) | 156 (25.2) | 819 (27.6) | 41,109 (27.9) | 16,266 (23.2) | 57,375 (26.4) |
| Other behavioural/neurological disorder ever, n (%) | 112 (4.8) | 20 (3.2) | 132 (4.5) | 8259 (5.6) | 2751 (3.9) | 11,010 (5.1) |
| Ever depression, n (%) | 930 (39.7) | 224 (36.1) | 1154 (38.9) | 60,347 (41.0) | 24,289 (34.7) | 84,636 (39.0) |
| Ever antidepressants, n (%) | 1205 (51.4) | 303 (48.9) | 1508 (50.9) | 73.778 (50.1) | 30,152 (43.1) | 103,930 (47.9) |
| Ever antipsychotics, n (%) | 501 (21.4) | 97 (15.7) | 598 (20.2) | 27,183 (18.5) | 9255 (13.2) | 36,438 (16.8) |
| Ever hypnotics/anxiolytics, n (%) | 1116 (47.6) | 267 (43.1) | 1383 (46.6) | 55,871 (38.0) | 23,535 (33.6) | 79,406 (36.6) |
| Ever mood stabilisers, n (%) | 149 (6.4) | 16 (2.6) | 165 (5.6) | 7089 (4.8) | 1706 (2.4) | 8795 (4.1) |
We extracted information on the number of patients in the entire CPRD who were diagnosed with cancer each year. We used this information to calculate the proportion of incident cancer cases who were current smokers. The proportion of patients with incident diagnoses of cancer was steady between 9.9% and 12.1% between 2007 and 2012 (Figure 15). Since 2012, the proportion of these patients who smoke has steadily dropped to 6.5% in 2015. In 2007, patients with a recent diagnosis of cancer were as likely to be prescribed NRT as those without a diagnosis, whereas they were slightly less likely to be prescribed varenicline. Prescribing of NRT fell over the period, but less quickly for those recently diagnosed with cancer (Figure 16). Prescribing of varenicline fell after 2012 for people with and without a recent cancer diagnosis. By 2015, patients with a cancer diagnosis were more likely to be prescribed varenicline than those without a cancer diagnosis.
FIGURE 15 The proportion of incident cancer patients in primary care who smoke, from 2007 to 2015. (Figure not reproduced here; see the original article.)
FIGURE 16 Prescription prevalence of varenicline or NRT in primary care, from 2007 to 2015, in smokers with any incident cancer diagnosis. (Figure not reproduced here; see the original article.)
Patients who were diagnosed with lung cancer were substantially more likely to be prescribed NRT than varenicline across the entire period; however, prescribing rates have fallen since 2010 (Figure 17). In contrast, prescribing rates for varenicline in this sample have increased from 2.4% in 2007 to 5.6% in 2015.
FIGURE 17 Prescription prevalence of varenicline or NRT in primary care, from 2007 to 2015, in smokers with an incident lung cancer diagnosis. (Figure not reproduced here; see the original article.)
Patients recently diagnosed with cancer were more likely to be abstinent 2 years after their first prescription than patients without a diagnosis (Figure 18). A total of 32.7% and 29.8% of those prescribed varenicline and NRT, respectively, were abstinent 2 years after their first prescription. However, these figures are much more imprecise than the equivalent figures for the whole sample. They are too imprecise to test whether or not there is a difference in effectiveness in this population compared with the whole population.
FIGURE 18 Percentage of patients with an electronic medical record indicating smoking cessation at 2-year follow-up, by exposure, and incident cancer status. (Figure not reproduced here; see the original article.)