A Cost- Benefit Analysis of Immunizing Healthy Adults Against Influenza

As of 11:30CST, 11/​11/​14, this cost-benefit anal­y­sis has been re­vised, in or­der to ad­dress con­cerns raised in the com­ments. See http://​​less­wrong.com/​​r/​​dis­cus­sion/​​lw/​​l8k/​​ex­pan­sion_on_a_pre­vi­ous_cost­benefit_anal­y­sis_of/​​ for more on how the cost-benefit anal­y­sis was car­ried out, and on how vary­ing cer­tain pa­ram­e­ters af­fected the de­ter­mined ex­pected value of re­ceiv­ing a flu shot.

Overview

The pur­pose of this post is to provide read­ers of LessWrong with a sum­mary of what the liter­a­ture has to say about the effi­cacy and safety of in­fluenza vac­ci­na­tions, as well as to weigh the costs of re­ceiv­ing yearly flu vac­ci­na­tions against the benefits which healthy adults gain from vac­ci­na­tion. As illus­trated in the “Cost-Benefit Analy­ses” sec­tion of this re­port, the ex­pected value of re­ceiv­ing flu vac­ci­na­tions is pos­i­tive for healthy adults. There­fore, a fur­ther mo­ti­va­tion for au­thor­ing this post is that writ­ing this post may en­courage LessWrong read­ers who have not yet been vac­ci­nated this flu sea­son to re­ceive im­me­di­ate vac­ci­na­tion.

In­tro­duc­tion and Re­view of Literature

Sev­eral meta-analy­ses on the effi­cacy and safety of live-at­ten­u­ated in­fluenza vac­cines, triva­lent in­ac­ti­vated in­fluenza vac­cines, and tetrava­lent in­ac­ti­vated in­fluenza vac­cines have been pub­lished within the last two years (see Cole­man et. al, Demicheli et. al, Oster­holm et. al). Th­ese meta- analy­ses reached broadly similar con­clu­sions re­gard­ing the effi­cacy of flu vac­cines, which groups were most at risk for be­ing in­fected with in­fluenza, the safety of be­ing vac­ci­nated, and the mag­ni­tude of so­cial harm caused yearly by in­fluenza. How­ever, there was dis­agree­ment be­tween some ar­ti­cles re­gard­ing whether or not vac­ci­na­tion of healthy adults against in­fluenza should be pur­sued as a pub­lic health policy. Speci­fi­cally, the Demicheli pa­per (wrongly) found “no ev­i­dence for the uti­liza­tion of vac­ci­na­tion against in­fluenza in healthy adults as a rou­tine pub­lic health mea­sure”. The is­sue of whether or not healthy adults should re­ceive flu shots will be ex­am­ined in the “Cost-Benefit Analy­ses” sec­tion of this re­port.

While the sever­ity of flu sea­sons varies greatly year-to-year, an av­er­age of 24,000 deaths from the flu oc­cur yearly in the US (NCIRD); ap­prox­i­mately 90% of these deaths are in peo­ple of at least 65 years of age (NCIRD, CDC Key Facts). For all flu sea­sons be­tween 1976 and 2007, an av­er­age of 2,385 adults of ages 19-64 died each year from flu and flu-re­lated causes in the US (Thomp­son). Between 5 and 20 per­cent of the US pop­u­la­tion be­comes in­fected with flu virus each flu sea­son (CDC Q&A).

The *effi­cacy* of a vac­cine is a mea­sure of how effec­tive a vac­cine is; if half of a pop­u­la­tion of 2,000,000 peo­ple were given a vac­cine with 60% effi­cacy, and 100,000 of the 1,000,000 to­tal un­vac­ci­nated peo­ple got sick, then 40,000 of the 1,000,000 vac­ci­nated peo­ple would get sick, as well. Many sources re­port the av­er­age effi­cacy of the flu vac­cine through­out the US pop­u­la­tion to be 60% (Demicheli et. al) or 59% (Oster­holm et. al, Cole­man et. al). The CDC re­ports that the flu vac­cine is more effi­ca­cious in young adults (70-90% effi­cacy, de­pend­ing on how closely ac­tive viruses match the ones in­cluded in the vac­cines man­u­fac­tured dur­ing a given sea­son), and less effi­ca­cious in those over 65 years of age (NCIRD). This has led to in­creased efforts at tar­get­ing health­care work­ers, nurs­ing home at­ten­dants, and oth­ers who are in fre­quent con­tact with el­derly per­sons for yearly vac­ci­na­tion.

While some health agen­cies only recom­mend that el­derly, in­fants, health­care work­ers, preg­nant women, and adults with cer­tain med­i­cal com­pli­ca­tions, such as res­pi­ra­tory dis­eases, re­ceive flu shots, the CDC recom­mends that all peo­ple 6 months and older get a flu shot ev­ery year (CDC Key Facts). The value which such at-risk in­di­vi­d­u­als gain from be­ing im­mu­nized against the flu is higher than the value which healthy adults gain from re­ceiv­ing flu shots. Cer­tain in­di­vi­d­u­als with ex­tremely rare con­di­tions, such as Guillain-Barré Syn­drome (GBS), or peo­ple who may ex­pe­rience life-threat­ing aller­gic re­ac­tions to com­po­nents of the flu shot, should not re­ceive flu shot. A health­care pro­fes­sional will be able to tell you whether or not it is safe for you to re­ceive a flu shot prior to you re­ceiv­ing the im­mu­niza­tion.

None of the meta-re­views ex­am­ined in this re­port found any ev­i­dence that re­ceiv­ing an in­fluenza vac­cine can cause se­ri­ous ad­verse re­sponses in pa­tients (Cole­man et. al, Oster­holm et. al, Demicheli et. al). Re­ceiv­ing the in­fluenza vac­cine is safe, and it is not at all pos­si­ble to catch the flu from re­ceiv­ing an in­fluenza vac­cine (CDC). Flu shots can cause arm pain or sore­ness, and can cause headache, mild fever, and mus­cle pain (Cole­man et. al, Demicheli et. al).

Cost-Benefit Analyses

Es­ti­mates of the ex­pected mon­e­tary val­ues of pos­si­ble flu-re­lated out­comes were calcu­lated rel­a­tive to the value of not get­ting sick de­spite not re­ceiv­ing a flu shot, which was defined as hav­ing a util­ity of 0 USD. All pay­offs shown are the pay­offs which an av­er­age in­di­vi­d­ual would de­rive from ex­pe­rienc­ing par­tic­u­lar out­comes, rather than the value which ei­ther so­ciety, em­ploy­ers, or other par­ties would gain from a given in­di­vi­d­ual ei­ther get­ting sick or not. Prob­a­bil­ities were as­signed to each out­come, as shown in Figure 1, and a calcu­la­tion of the ex­pected value of re­ceiv­ing or not re­ceiv­ing a flu shot in a given year was car­ried out. The mo­ti­va­tion for sim­plify­ing the calcu­la­tion of the ex­pected value of re­ceiv­ing a flu shot by re­strict­ing the out­come space as shown in Figure 1 was to demon­strate that, de­spite us­ing con­ser­va­tive es­ti­mates and ig­nor­ing cer­tain benefits of vac­ci­na­tion in the model, the ex­pected value of vac­ci­na­tion is still pos­i­tive for healthy adults. Since other de­mo­graph­ics are ex­pected to benefit even more from re­ceiv­ing flu vac­ci­na­tions than healthy adults benefit from re­ceiv­ing flu vac­ci­na­tions, the fact that healthy adults would benefit from re­ceiv­ing yearly flu vac­ci­na­tions strongly sug­gests that all in­di­vi­d­u­als above 6 months of age would benefit from re­ceiv­ing flu shots, ex­cept­ing e.g. pa­tients with GBS or aller­gies to com­po­nents of the flu shot.

The cost of get­ting a flu shot was calcu­lated as be­ing 30 USD, given that it costs around 20 USD to re­ceive a flu shot out of pocket, and given that it takes around 30 min­utes to get a flu shot at a clinic. I have es­ti­mated the value of one’s time as be­ing 20 USD/​hour for this calcu­la­tion.

The value of not feel­ing sick for 3-10 days was sub­jec­tively es­ti­mated as be­ing 200 USD for those who caught the flu, yet did not re­ceive a flu shot. The out­come in which one catches the flu de­spite re­ceiv­ing a flu shot was given a pay­off of − 230 USD, which was calcu­lated by adding the cost of be­ing vac­ci­nated against the flu to the cost of feel­ing sick from get­ting the flu, calcu­lated above.

The value a given in­di­vi­d­ual would gain from not dy­ing was es­ti­mated as be­ing 5,000,000 USD.

Figure 1. De­ci­sion Tree for Assess­ing the Im­pact of Im­mu­niza­tion in Healthy Adults

Although the costs of the pos­si­ble out­comes shown in Figure 1 were calcu­lated un­der the as­sump­tion that the in­di­vi­d­ual re­ceiv­ing the flu shout was un­in­sured, hav­ing an in­surance policy in­creases the ex­pected value of re­ceiv­ing a flu shot, as many in­surance com­pa­nies will com­pletely cover the cost of re­ceiv­ing a flu shot. Some gov­ern­men­tal health in­surance pro­grams do not cover the cost of flu vac­ci­na­tions. If one has in­surance which cov­ers the cost of the flu vac­cine, the ex­pected value of be­ing vac­ci­nated against the flu rises by 20 USD.

There are sev­eral pos­i­tive benefits of re­ceiv­ing flu shots which have not been in­cluded in the above model. In par­tic­u­lar, be­ing vac­ci­nated against the flu pro­tects oth­ers in your com­mu­nity from be­com­ing sick; this effect is known as the herd im­mu­nity effect. Also, the above anal­y­sis as­sumed that an in­di­vi­d­ual would not lose in­come from miss­ing work due to be­ing sick from the flu; the effect which mak­ing this as­sump­tion had on the cost-benefit anal­y­sis pre­sented here is ex­am­ined in the link given at in the first para­graph of this post. Lastly, re­ceiv­ing the flu vac­cine pro­vides one with a small de­gree of pro­tec­tion against in­fluenza-like in­fec­tions (Cole­man et. al, Demicheli et. al); this pos­i­tive effect of the flu vac­cine was not con­sid­ered in the above as­sess­ment of the costs and benefits as­so­ci­ated with healthy adults re­ceiv­ing the flu vac­cine.

Again, the above anal­y­sis of the ex­pected util­ity of re­ceiv­ing the flu vac­cine each year was con­ducted with con­ser­va­tive es­ti­mates and a sim­ple model which did not take into ac­count all of the benefits of re­ceiv­ing the flu vac­cine; this was done to show that the ex­pected gain from re­ceiv­ing a flu shot is pos­i­tive in the gen­eral case, given un­char­i­ta­ble as­sump­tions.

Author’s Reflections

I only read the “Meth­ods”, “Find­ings”, and “In­ter­pre­ta­tion” sec­tions of the Lancet ar­ti­cle, as I did not have ac­cess to the full text of this pa­per.

Be­fore writ­ing this ar­ti­cle and con­duct­ing the re­search which nec­es­sar­ily had to be con­ducted be­fore writ­ing it, I would have es­ti­mated the prior prob­a­bil­ity of el­derly peo­ple, in­fants, preg­nant woman, and asth­mat­ics re­ceiv­ing a net benefit from vac­ci­na­tion as be­ing very high, and the prior prob­a­bil­ity of healthy adults re­ceiv­ing a net benefit from in­fluenza vac­ci­na­tion as mod­er­ately high.

I was raised in a fam­ily which, in gen­eral, val­ued be­ing healthy, and, in par­tic­u­lar, val­ued the prac­tice of keep­ing up to date on one’s vac­ci­na­tions. How­ever, I do not be­lieve that the con­clu­sions of this re­port would have been differ­ent if I had not come from such a cul­ture.

Fur­ther Considerations

While this re­port is com­plete, I could have been more thor­ough. Part of why I am pub­lish­ing this post now, rather than con­duct­ing more re­search be­fore do­ing so, is that I ex­pect that con­duct­ing ad­di­tional re­search would be very un­likely to cause me to change any of the ma­jor con­clu­sions of this re­port. To say the same thing from a de­ci­sion-the­o­retic stand­point, in­for­ma­tion which has a very low chance of mak­ing one change their mind about some­thing has lit­tle value, and I think that read­ing more pa­pers on this topic would have a very low chance of chang­ing any of my opinions on this topic.

References

1. Cen­ters for Disease Con­trol and Preven­tion. Key Facts About Sea­sonal Flu Vac­cine. http://​​www.cdc.gov/​​flu/​​pro­tect/​​key­facts.htm (ac­cessed 119, 2014).

2. Cen­ters for Disease Con­trol and Preven­tion. Sea­sonal In­fluenza Q&A. http://​​www.cdc.gov/​​flu/​​about/​​qa/​​dis­ease.htm (ac­cessed 119, 2014).

3. Cole­man, B.; Cochrane, L.; Co­las, L. Liter­a­ture Re­view on Quadri­va­lent In­fluenza Vac­cines. Public Health Agency of Canada 2014.

4. Demicheli, V.; Jeffer­son, T.; Al-An­sary, L.; Fer­roni, E. Vac­cines for pre­vent­ing in­fluenza in healthy adults. Cochrane Library 2014.

5. Milenkovic, M.; Russo, A.; Elix­hauser, A. Hospi­tal Stays for In­fluenza, 2004. Agency for Health­care Re­search and Qual­ity 2006.

6. Na­tional Cen­ter for Im­mu­niza­tion and Re­s­pi­ra­tory Diseases. Epi­demiol­ogy and Preven­tion of Vac­cine-Preventable Diseases. http://​​www.cdc.gov/​​vac­cines/​​pubs/​​pink­book/​​flu.html (ac­cessed 119, 2014).

7. Oster­holm, M. T.; Kel­ley, N. S.; Som­mer, A.; Belon­gia, E. A. Effi­cacy and effec­tive­ness of in­fluenza vac­cines: a sys­tem­atic re­view and meta-anal­y­sis. The Lancet in­fec­tious dis­eases 2012, 12, 36-44.

8. Thomp­son, M.; Shay, D.; Zhou, H.; Bridges, C. Es­ti­mates of Deaths As­so­ci­ated with Sea­sonal In­fluenza—United States, 1976--2007. 2010.