The impact of firms’ adjustments on the indirect cost of illness

Abstract

Illness-related absenteeism reduces firms’ output, an effect referred to as indirect cost (IC) and often included in cost-of-illness or cost-effectiveness (of health technologies) studies. The companies may foresee this effect and modify hiring or contracting policies. We present a model allowing the estimation of IC with such adjustments. We show that the risk of illness does not change the general shape and properties of the (expected) marginal productivity function. We apply our model to several illustrative examples and show that firm’s adjustments impact IC in an ambiguous way, depending on detailed company/market characteristics: in some cases the company reduces the employment (further increasing IC), in another—the opposite happens. Contrary to previous findings, teamwork and shortfall penalties may reduce IC in some settings. Our analysis highlights that IC should be split into the result of companies preparing for and actually experiencing sick leaves.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Notes

  1. 1.

    For example, in Austria (Walter and Zehetmayr 2006), Sweden (Pharmaceutical Benefits Board 2003), the Netherlands (College Voor Zorgverzekeringen 2006), Australia (Pharmaceutical Benefits Advisory Committee 2008), or Italy (Capri et al. 2001).

  2. 2.

    This assumption approximately holds, e.g., in Poland, where the employer pays the salary for the first 33 days of the illness, and then the public insurer takes over (the employee gets 80% of a regular wage). Similar model is used in Germany, where the employer pays full salary for at least first 6 weeks of illness, then the insurer pays 70% of salary. In Sweden, employer covers sick-leaves for the first 2 weeks of illness.

References

  1. Arnold, D., & de Pinto, M. (2014). Sickness absence, presenteeism and work-related characteristics: Conference paper. Conference: Beiträge zur Jahrestagung des Vereins für Socialpolitik 2015: Ökonomische Entwicklung—Theorie und Politik—Session: Labor-Empirical Studies 5, No G16-V1.

  2. Audrey, S., & Procter, S. (2015). Employers’ views of promoting walking to work: A qualitative study. International Journal of Behavioral Nutrition and Physical Activity. doi:10.1186/s12,966-015-0174-8.

  3. Berger, M., Murray, J., Xu, J., & Pauly, M. (2001). Alternative valuations of work loss and productivity. Journal of Occupational and Environmental Medicine, 43, 18–24.

    CAS  Article  PubMed  Google Scholar 

  4. Brouwer, W., van Exel, N., Baltussen, R., & Rutten, F. (2006). A dollar is a dollar is a dollar—Or is it? Value in Health, 9, 341–347.

    Article  PubMed  Google Scholar 

  5. Bryson, A. (2013). Do temporary agency workers affect workplace performance? Journal of Productivity Analysis, 39, 131–138.

    Article  Google Scholar 

  6. Burton, W., Morrison, A., & Wertheimer, A. (2003). Pharmaceuticals and worker productivity loss: A critical review of the literature. Journal of Occupational and Environmental Medicine, 45, 610–621.

    Article  PubMed  Google Scholar 

  7. Capri, S., Ceci, A., Terranova, L., Merlo, F., & Mantovani, L. (2001). Guidelines for economic evaluations in Italy: Recommendations from the Italian group of pharmacoeconomic studies. http://www.ispor.org/PEguidelines/source/Italy%20PE%20guidelines_Italy.

  8. College Voor Zorgverzekeringen (2006). Guidelines for pharmacoeconomic research. http://www.ispor.org/PEguidelines/source/HTAGuidelinesNLupdated2006.

  9. DeLeire, T., & Manning, W. (2004). Labor market costs of illness: Prevalence matters. Health Economics, 13, 239–250.

    Article  PubMed  Google Scholar 

  10. Knies, S., Boonen, A., Candel, M., Evers, S., & Severens, J. (2013). Compensation mechanisms for lost productivity: A comparison between four European countries. Value in Health, 16, 740–744.

    CAS  Article  PubMed  Google Scholar 

  11. Koopmanschap, M., Rutten, F., van Ineveld, B., & van Roijen, L. (1995). The friction cost method for measuring indirect costs of disease. Journal of Health Economics, 14, 171–189.

    CAS  Article  PubMed  Google Scholar 

  12. Koopmanschap, M., & van Ineveld, B. (1992). Towards a new approach for estimating indirect cost of disease. Social Science and Medicine, 34, 1005–1010.

    CAS  Article  PubMed  Google Scholar 

  13. Krol, M., Brouwer, W., & Rutten, F. (2013). Productivity costs in economic evaluations: Past, present, future. Pharmacoeconomics, 31, 537–549.

    Article  PubMed  Google Scholar 

  14. Krol, M., Brouwer, W., Severens, J., Kaper, J., & Evers, S. (2012). Productivity cost calculations in health economic evaluations: Correcting for compensation mechanisms and multiplier effects. Social Science and Medicine, 75, 1981–1988.

    Article  PubMed  Google Scholar 

  15. Leigh, J. P. (1981). The effects of union membership on absence from work due to illness. Journal of Labor Research, 2, 329–336.

    Article  Google Scholar 

  16. Liljas, B. (1998). How to calculate indirect costs in economic evaluations. Pharmacoeconomics, 13, 1–17.

    CAS  Article  PubMed  Google Scholar 

  17. Loeppke, R., Hohn, T., Baase, C., Bunn, W., Burton, W., Eisenberg, B., et al. (2015). Integrating health and safety in the workplace: How closely aligning health and safety strategies can yield measurable benefits. Journal of Occupational and Environmental Medicine, 57, 585–597.

    Article  PubMed  Google Scholar 

  18. Pauly, M., Nicholson, S., Polsky, D., Berger, M., & Sharda, C. (2008). Valuing reductions in on-the-job illness: ‘Presenteeism’ from managerial and economic perspectives. Health Economics, 17, 469–485.

    Article  PubMed  Google Scholar 

  19. Pauly, M., Nicholson, S., Xu, P. D. J., Danzon, P., Murray, J., & Berger, M. (2002). A general model of the impact of absenteeism on employers and employees. Health Economics, 11, 221–231.

    Article  PubMed  Google Scholar 

  20. Pharmaceutical Benefits Advisory Committee (2008). Guidelines for preparing submissions to the pharmaceutical benefits advisory committee. http://www.ispor.org/PEguidelines/source/Australia-Guidelines-for-preparing-submissions-to-the-Pharmaceutical-Benefits-Advisory-Committee-2008.

  21. Pharmaceutical Benefits Board (2003). General guidelines for economic evaluations from the pharmaceutical benefits board. http://www.ispor.org/PEguidelines/source/Guidelines_in_Sweden.

  22. Suhrcke, M., Sauto Arce, R., McKee, M., & Rocco, L. (2012). Economic costs of ill health in the European region. In J. Figueras & M. McKee (Eds.), Health systems, health, wealth and societal well-being: Assessing the case for investing in health systems (European Observatory on Health Care Systems). New York: Mcgraw Hill.

    Google Scholar 

  23. Walter, W., & Zehetmayr, S. (2006). Guidelines on health economic evaluation:consensus paper. http://www.ispor.org/PEguidelines/source/Guidelines_Austria.

Download references

Acknowledgements

This paper was partially prepared when M. Jakubczyk was at The University of Iowa, Tippie College of Business, which was possible thanks to the Fulbright Senior Award 2015/2016. We would also like to thank R. Amir for useful comments.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Michał Jakubczyk.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Appendix

Appendix

Proof of Proposition 1

We first show how \(\text {MP}^E({\cdot })\) can be presented in a different form, more convenient for the present proof. We start with rewriting Eq. 4 and then modify it (notice we write it for \((L+1)\)-th worker here and change how the summation range is denoted)

$$\begin{aligned} \begin{aligned}&\frac{\text {MP}^E(L+1)}{1-s}=\sum _{i=0}^L {L \atopwithdelims ()i} s^{L-i}(1-s)^i \text {MP}(i+1)\\&\quad =\sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i} s^{L-i}(1-s)^i\text {MP}(i+1) + \sum _{i=1}^L {L-1 \atopwithdelims ()i-1} s^{L-i}(1-s)^i\text {MP}(i+1)\\&\quad =s \sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i} s^{L-1-i}(1-s)^i\text {MP}(i+1)\\&\qquad + (1-s) \sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i} s^{L-1-i}(1-s)^i\text {MP}(i+2)\\&\quad =s\times \frac{\text {MP}^E(L)}{1-s} + (1-s) \sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i} s^{L-1-i}(1-s)^i\text {MP}(i+2), \end{aligned} \end{aligned}$$
(8)

where we use the fact that for \(i\in {\mathbf {N}}, 0<i<L, {L \atopwithdelims ()i}={L-1 \atopwithdelims ()i} + {L-1 \atopwithdelims ()i-1}\). Hence, \(\text {MP}^E(L+1)\) is a weighted average (with weights s and \(1-s\)) of \(\text {MP}^E(L)\) and a similarly calculated expression in which we take \(\text {MP}\)s moved one worker to the right. As \(\text {MP}({\cdot })\) is decreasing it must be that

$$\begin{aligned} \sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i} s^{L-1-i}(1-s)^i\text {MP}(i+1)\ge \sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i} s^{L-1-i}(1-s)^i\text {MP}(i+2), \end{aligned}$$

as \(\ge \) holds for each term of the summation. \(\square \)

Proof of Proposition 2

The proof takes several steps. First, notice that \(\text {MP}^E(1)=(1-s)\text {MP}(1)\), and \(\text {MP}^E(2)=(1-s)\left( s\times \text {MP}(1)+(1-s)\times \text {MP}(2)\right) \). Then, if \(\text {MP}(2)\ge \text {MP}(1)\), then \(\text {MP}^E(2)\ge \text {MP}^E(1)\) for any s. Thus, if \(\text {MP}({\cdot })\) is increasing initially, so is \(\text {MP}^E({\cdot })\).

Notice, that if \(\text {MP}({\cdot })\) is decreasing after some point, then \(\text {MP}^E\) must also decrease at some point. Say, we calculate \(\text {MP}^E({\cdot })\) for some \(0\ll L_1\ll L_2\), where already \(L_1\) is much greater than the point where \(\text {MP}({\cdot })\) started to be decreasing. Increasing \(L_2\) results in \(\text {MP}^E(L_2\)) being calculated (as a weighted sum in Eq. 4) averaging lower values of \(\text {MP}({\cdot })\) than when calculating \(\text {MP}^E(L_1)\) (the weights assigned to the values of \(\text {MP}({\cdot })\) prevailing in both sums can be made arbitrarily small by increasing \(L_2\)).

We now show that if \(\text {MP}^E(L+1)<\text {MP}^E(L)\) then \(\text {MP}^E({\cdot })\) must be strictly decreasing further on. This is the main part of the whole proof, and again it is done in several steps. First, \(\text {MP}^E({\cdot })\) is calculated as an expected value of \(\text {MP}({\cdot })\) function calculated for a variable distributed according to a binomial distribution. It is then useful to see how the probabilities change in this distribution. Denote by B(knp) the probability of k successes in n independent experiments, where a single success comes with probability p. We are interested in

$$\begin{aligned} \frac{B(k,n,p)}{B(k+1,n,p)}=\frac{{n \atopwithdelims ()k}p^k(1-p)^{n-k}}{{n\atopwithdelims ()k+1}p^{k+1}(1-p)^{n-k-1}}=\frac{k+1}{n-k}\times \frac{1-p}{p}, \end{aligned}$$
(9)

and this is increasing in k.

As noticed in the proof of Proposition 1, \(\text {MP}^E(L+1)\) is calculated as a weighted average of \(\text {MP}^E(L)\) (calculated based on \(\text {MP}({\cdot })\) for workers from 1 to L) and a weighted average of \(\text {MP}\) calculated for workers from 2 to \((L+1)\) with weights given by a binomial distribution \(B({\cdot },L-1,1-s)\). It will be useful to show that if

$$\begin{aligned} \sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+1)>\sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+2), \end{aligned}$$
(10)

and so \(\text {MP}^E(L+1)<\text {MP}^E(L)\), then also

$$\begin{aligned} \sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+2)>\sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+3). \end{aligned}$$
(11)

Let \({\varDelta }(i)\) denote \(\text {MP}(i)-\text {MP}(i-1)\), defined for \(i\ge 2\). \({\varDelta }({\cdot })\) is positive for the part of \(\text {MP}({\cdot })\) with increasing marginal productivity. Our assumptions guarantee that \({\varDelta }({\cdot })\) changes from positive to negative. Rewriting inequality 10

$$\begin{aligned} \sum _{i=0}^{L-1} B(i,L-1,1-s){\varDelta }(i+2)<0, \end{aligned}$$
(12)

and, analogously, in order to prove inequality 11 we need to show that

$$\begin{aligned} \sum _{i=0}^{L-1} B(i,L-1,1-s){\varDelta }(i+3)<0. \end{aligned}$$
(13)

Obviously some \({\varDelta }({\cdot })\) values in inequality 12 must be negative. Assume non-trivially that also some (for \(i\le i^*\)) are positive. Then

$$\begin{aligned} \begin{aligned}&\sum _{i=0}^{L-1} B(i,L-1,1-s){\varDelta }(i+2) \ge \sum _{i=1}^{L-1} B(i,L-1,1-s){\varDelta }(i+2)\\&\quad =\underbrace{\sum _{i=1}^{i^*} B(i,L-1,1-s){\varDelta }(i+2)}_{\text {(I)}} + \underbrace{\sum _{i=i^*+1}^{L-1} B(i,L-1,1-s){\varDelta }(i+2)}_{\text {(II)}}, \end{aligned} \end{aligned}$$

where (I) is positive, and (II) is negative. We have \(\text {(I)}+\text {(II)}<0\), and so

$$\begin{aligned} \sum _{i=1}^{i^*} B(i-1,L-1,1-s){\varDelta }(i+2) + \sum _{i=i^*+1}^{L-1} B(i-1,L-1,1-s){\varDelta }(i+2)<0, \end{aligned}$$

as Eq. 9 is increasing in k, and so the negative elements in (II) are inflated more than the positive elements in (I). Rewriting the last inequality we get

$$\begin{aligned} \begin{aligned}&\sum _{i=1}^{L-1} B(i-1,L-1,1-s){\varDelta }(i+2)= \sum _{i=0}^{L-2} B(i,L-1,1-s){\varDelta }(i+3)\\&\quad \ge \sum _{i=0}^{L-1} B(i,L-1,1-s){\varDelta }(i+3), \end{aligned} \end{aligned}$$

which proves inequality 13. As the final step, notice that just as we decomposed \(\text {MP}^E(L+1)\), we can decompose \(\text {MP}^E(L+2)\):

$$\begin{aligned} \begin{aligned} \frac{\text {MP}^E(L+2)}{1-s}&=s^2 \sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+1)\\&\quad +2s(1-s) \sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+2)\\&\quad +(1-s)^2 \sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+3)\\&=s^2 \sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+1)\\&\quad +2s(1-s) \sum _{i=0}^{L-1} B(i,L-1,1-s) (\text {MP}(i+1)+{\varDelta }(i+2))\\&\quad +(1-s)^2 \sum _{i=0}^{L-1} B(i,L-1,1-s) (\text {MP}(i+1)+{\varDelta }(i+2)+{\varDelta }(i+3))\\&=\sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+1) +(1-s^2)\sum _{i=0}^{L-1} B(i,L-1,1-s) {\varDelta }(i+2)\\&\quad +(1-s)^2\sum _{i=0}^{L-1} B(i,L-1,1-s) {\varDelta }(i+3)<\sum _{i=0}^{L-1} B(i,L-1,1-s) \text {MP}(i+1) \\&\quad +(1-s)\sum _{i=0}^{L-1} B(i,L-1,1-s) {\varDelta }(i+2)=\frac{\text {MP}^E(L+1)}{1-s}, \end{aligned} \end{aligned}$$
(14)

where the inequality results from omitting a negative term and \(1-s^2\ge 1-s\). This finishes the proof of \(\text {MP}^E({\cdot })\) being decreasing ever since the first decrease.

Finally, to show that the first decrease will only happen for some L greater than the point at which \(\text {MP}({\cdot })\) attains its maximum, it’s enough to revert the argument presented in the proof of proposition 1. \(\square \)

Proof of Proposition 3

We will show that if for some L we have \(\text {MP}^E(L) \le \text {MP}^E(L+1) \ge \text {MP}^E(L+2)\), then the difference \(\text {MP}^E(L+2)-\text {MP}^E(L+1)\) is strictly increasing in s. Based on the decomposition presented in the proof of Proposition 1, this difference amounts to

$$\begin{aligned} s\times \text {MP}^E(L+1)+(1-s)^2\sum _{i=0}^L {L \atopwithdelims ()i}(1-s)^i s^{L-i}\text {MP}(i+2)-\text {MP}^E(L+1), \end{aligned}$$

and so to

$$\begin{aligned} (1-s)^2\sum _{i=0}^L {L \atopwithdelims ()i}(1-s)^i s^{L-i}\text {MP}(i+2)-(1-s)^2\sum _{i=0}^L {L \atopwithdelims ()i}(1-s)^i s^{L-i}\text {MP}(i+1) \end{aligned}$$

or, more concisely, to \((1-s)^2\sum _{i=0}^L {L \atopwithdelims ()i}(1-s)^i s^{L-i}{\varDelta }(i+2)\).

The derivative of this expression with respect to s is given as \(\text {(I)}+\text {(II)}+\text {(III)}+\text {(IV)}+\text {(V)}\), where

$$\begin{aligned} \text {(I)}&=-2(1-s)\sum _{i=0}^L {L \atopwithdelims ()i}(1-s)^i s^{L-i}{\varDelta }(i+2),\\ \text {(II)}&=(1-s)^2{L \atopwithdelims ()0}Ls^{L-1}{\varDelta }(2),\\ \text {(III)}&=-(1-s)^2\sum _{i=1}^{L-1}\frac{L!}{(i-1)!\,(L-i)!}(1-s)^{i-1}s^{L-i}{\varDelta }(i+2),\\ \text {(IV)}&=(1-s)^2\sum _{i=1}^{L-1}\frac{L!}{i!\,(L-i-1)!}(1-s)^i s^{L-i-1}{\varDelta }(i+2),\\ \text {(V)}&=-(1-s)^2{L \atopwithdelims ()L}L(1-s)^{L-1}{\varDelta }(L+2). \end{aligned}$$

Now, \(\text {(II)}+\text {(IV)}=(1-s)^2 L\sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i}(1-s)^i s^{L-1-i}{\varDelta }(i+2)\), while \(\text {(III)}+\text {(V)}=-(1-s)^2 L\sum _{i=0}^{L-1} {L-1 \atopwithdelims ()i}(1-s)^i s^{L-1-i}{\varDelta }(i+3)\).

Notice (cf. Eq. 8 in the proof of Proposition 1) that

$$\begin{aligned} \text {MP}^E(L+1)-\text {MP}^E(L)=(1-s)^2\sum _{i=0}^{L-1}{L-1 \atopwithdelims ()i}s^{L-1-i}(1-s)^i{\varDelta }(i+2). \end{aligned}$$
(15)

Accounting for our assumptions, that directly proves that \(\text {(II)}+\text {(IV)}\ge 0\) and \(\text {(I)}\ge 0\) (applying the formula to \(\text {MP}^E(L+2)-\text {MP}^E(L+1)\)).

We can also decompose \(\text {MP}^E(L+2)\) into three elements as shown in first three lines of Eq. 14 in the proof or Proposition 2. Again, accounting for the fact that \(\text {MP}^E(L) \le \text {MP}^E(L+1) \ge \text {MP}^E(L+2)\) it must be that \(\text {(III)}+\text {(V)}\ge 0\).

If either \(\text {MP}^E(L) < \text {MP}^E(L+1)\) or \(\text {MP}^E(L+1) > \text {MP}^E(L+2)\), then the derivative is strictly positive (for \(s<1\)). \(\square \)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Jakubczyk, M., Koń, B. The impact of firms’ adjustments on the indirect cost of illness. Int J Health Econ Manag. 17, 377–394 (2017). https://doi.org/10.1007/s10754-017-9212-1

Download citation

Keywords

  • Absenteeism
  • Indirect cost
  • Teamwork
  • Output shortfall
  • Friction cost method
  • Societal perspective

JEL Classification

  • D21
  • J24
  • J21
  • L23