Smith, Kenneth J. MD, MS*; Tsevat, Joel MD, MPH†; Ness, Roberta B. MD, MPH‡; Wiesenfeld, Harold C. MD, CM§; Roberts, Mark S. MD, MPP*
PELVIC INFLAMMATORY DISEASE (PID) is a common, potentially serious, and expensive disease of young women.1,2 In the United States, it is estimated that 10% of women will be affected by PID during their lifetime, with aggregate costs of approximately $2 billion per year.3 PID is typically caused by sexually transmitted infections, most commonly chlamydia or gonorrhea.1 PID is rarely fatal, thus its impact is mainly through morbidity and decreases in quality of life that occur due to acute PID and its sequelae.
In cost-effectiveness analyses, the impact on quality of life is quantified through the use of utility values.4,5 Utilities are measured on a scale from 0 (dead) to 1 (perfect health) and denote one’s preference or desirability for a health state or condition.6,7 One such method is the time tradeoff (TTO), which measures preferences for a shorter but healthier life. Quality-adjusted life years, the recommended measure of effectiveness in cost-effectiveness analyses, are then calculated by multiplying the years of life in a health state by the utility of the health state.6 In published cost-effectiveness analyses of interventions to prevent or treat PID, utilities for PID-associated health states have typically come from sources other than women with PID or at risk for PID.8–11 The Panel on Cost-Effectiveness in Health and Medicine recommends using utilities elicited from persons in the community to reflect societal preferences in their recommended “reference case” societal perspective analysis.4 Utilities from women with PID are needed for analyses from the patient perspective. In addition, the US Preventive Services Task Force has called for analyses from the societal perspective as an aid in making their recommendations.12 In this study, we elicited quality of life utilities from women from the community with and without a history of PID, using a validated computerized questionnaire.
We recruited women older than 18 from Pittsburgh in 2005–2006. Women with a self-reported history of PID who participated in the PID Evaluation and Clinical Health (PEACH) study,13,14 a multicenter trial of PID treatment, were recruited by letter, and women attending the Allegheny County Sexually Transmitted Disease Clinic or the Magee Women’s Hospital emergency department were recruited in person. Women without a PID history were recruited from the same county STD clinic. The University of Pittsburgh Institutional Review Board approved the study protocol.
After obtaining written informed consent, subjects completed a computerized questionnaire with the help of a research assistant using Impact 3.0, a standard utility elicitation software program.15 First, we obtained demographic, childbearing, and PID symptom-related information (Appendix 1). We then administered the SF-12 health status instrument, a shortened, validated version of the SF-36.16 The SF-12 yields 2 summary scores, the Physical Component Summary and the Mental Component Summary, each of which is normed to 50 with a standard deviation of 10.
Next, we obtained visual analogue scale (VAS) and TTO valuations for 5 PID-associated health states: ambulatory PID treatment, hospital PID treatment, ectopic pregnancy, chronic pelvic pain, and infertility. Subjects read brief scenarios describing each state; then, using the computerized software as a visual aid, they gave VAS and TTO valuations (in that order) for the PID health states in the order listed above (i.e., in fixed order). In each scenario we asked subjects to consider a 25-year-old woman with that condition and included a description in lay person’s terms of the condition, information about the symptoms associated with the condition, treatment options, complication risks, and functional limitations that might occur as a result of the condition (Appendix 2). Where appropriate we used terminology from the SF-12 to describe symptoms and limitations that might occur as a result of these health states. Based on the scenarios, the PID treatment states are considered short-term health states, the ectopic pregnancy state is a short-term state with a possibility of long-term consequences (infertility and recurrent ectopic pregnancy), and infertility and chronic pelvic pain are long-term health states. These scenarios were reviewed by PID experts to ensure clinical fidelity, and modified based on their recommendations.
In the VAS, subjects were asked to indicate where individual health states fell on a vertical scale anchored by dead at the bottom (valued at 0%) and perfect health at the top (valued at 100%). In the TTO valuation, subjects choose between living a fixed time (10 years) in the health state being evaluated, or varying amounts of time in perfect health. The time in perfect health was varied in “ping-pong” fashion, until the point of indifference between choices was reached; the utility was calculated as the time in perfect health divided by the time in the given health state. Lower scores by either method indicate lower ratings or valuations for those states. If subjects rated ambulatory PID treatment worse than other states, they had the opportunity to modify (or, as in previously used terminology,17 “repair”) their valuations by raising their ambulatory treatment valuation, lowering the valuation for the given health state, or keeping their valuations unchanged. All utility assessments were performed by the same research assistant. Each interview lasted from 20 to 30 minutes, and subjects were compensated $25 for their time.
Demographic, childbearing, and symptom characteristics as well as SF-12 scores and utilities were compared between women with versus without a history of PID by using χ2 testing for categorical variables and the nonparametric Mann-Whitney U test for continuous variables. We used multiple linear regression to control for differences in characteristics between women with versus without a history of PID in comparing their utilities. A priori power calculations showed that with 60 patients with a history of PID and 150 controls there would be 80% power to detect a difference in utility of 0.0645 between groups. The larger number of controls was chosen to increase the power to detect differences between groups, taking into consideration also, based on PEACH study experience, that women with a PID history would be more difficult to recruit. With 56 patients with PID, the detectable difference increases to 0.0661, still within the range of the minimally important difference for utilities.18
We obtained data from 56 women with a history of PID [12 (21.4%) from the PEACH study and the remainder recruited from the other facilities] and 150 women with no PID history (Table 1). Women with prior PID were on average more than 4 years older and were significantly more likely to be nonwhite, unemployed, and to have children. Women with a PID history also had lower SF-12 Physical and Mental Health scores. No differences in desire for future fertility were seen between groups.
VAS valuations showed significant differences between patients with prior PID versus no prior history of PID with regards to values for ectopic pregnancy and chronic pelvic pain (P <0.05), and for infertility (P <0.01; Table 2). For example, the mean (standard deviation) VAS score for infertility was 0.59 (0.23) in women with prior PID, as opposed to 0.66 (0.24) among women who had not had PID (P = 0.003). Women with a PID history also tended to value those 3 states lower by the TTO method; however, those differences did not reach statistical significance. Valuations were similar between groups for ambulatory and inpatient treatment of PID by either method.
Differences in VAS valuations for ectopic pregnancy, pelvic pain, and infertility remained statistically significant between women with or without a PID history when controlling for age, race, employment, number of children, and importance of fertility (model R2’s 0.053–0.097). When also adjusting for SF-12 Physical and Mental Component Summary scores, between-group differences remained significant for chronic pelvic pain and infertility, but not for ectopic pregnancy (P = 0.13) (model R2’s 0.104–0.132). Regression coefficients were not significant for variables other than PID history, except for SF-12 Physical Component Summary scores, where coefficients ranged from 0.69 to 0.78 (P <0.01).
Women with infertility rated the infertility state significantly lower than women without infertility by the VAS (0.43 vs. 0.63, P = 0.033) but not by the TTO (0.75 vs. 0.82, P = 0.49). Women with chronic pelvic pain valued the chronic pain state less by the TTO (0.59 vs. 0.79, P = 0.026) but not by VAS (0.48 vs. 0.51, P = 0.52). Women who did and did not want children or more children and those with or without a history of ectopic pregnancy rated their corresponding health states similarly by either method.
We found that disutilities for PID health states are relatively large and that women who have experienced PID tend to have lower valuations for certain PID-related health states, specifically ectopic pregnancy, chronic pelvic pain, and infertility, all of which are common complications of PID. Statistically significant differences between patients with prior PID and patients without prior PID were seen only on the VAS, not the TTO, although absolute mean differences between groups for these health states were similar for each valuation method. Such differences in utilities for PID health states could affect outcomes of cost-effectiveness analyses conducted from the societal perspective as compared with the patient perspective. For example, lower utilities for PID complications would lead to greater effectiveness and more favorable cost-effectiveness ratios for interventions to prevent or treat PID in analyses performed from perspective of women with a PID history compared with analyses taking a societal perspective. VAS valuations in this study are similar to utilities derived from the Health Utility Index by an Institute of Medicine panel10 for outpatient PID treatment (0.63), inpatient treatment without surgery (0.57), ectopic pregnancy (0.58), and pelvic pain (0.60); their utility for infertility (0.82) is higher than our VAS values (0.53–0.66) but similar to our subjects’ TTO values for infertility (0.76–0.84). In general, however, VAS scores are lower than those obtained by other measures because the VAS does not involve valuation against an external metric, such as time or risk of death; as such, the VAS is less well accepted for decision analysis and cost-effectiveness analysis.7,19–22 TTO scores have a stronger theoretical basis and are generally accepted as “true” utilities.20,21,23
Two other studies also give insights into PID-related utilities. Gold et al.,5 using the HALex instrument derived from the National Health Interview survey, reported a utility of 0.75 for PID but did not assess utilities for specific aspects of the disease. In the PEACH study,24 women with PID compared the severity of infertility vis-à-vis a variety of non–PID-related health states. Results showed that infertility and chronic headaches were almost equally as likely to be considered “worse” than each other; utilities for chronic headaches have previously been found to range from 0.69 to 0.82.5,25,26 PEACH study subjects preferred infertility over paralysis (utility 0.42–0.45), but did not prefer infertility over sinus congestion (utility 0.84–1.0).5,24,25,27
Persons with a given condition will often value that condition more highly than persons without the condition due, at least in part, to adapting to the condition and reprioritizing health versus other aspects of life.28,29 Contrary to this effect, women in our study who had experienced PID tended to rate health states related to PID complications lower than women who had not had PID. Two possible explanations are (a) that unlike persons with other conditions, women with PID have not adapted to their health state, or (b) that our health state descriptions did not adequately capture the effects of PID complications on quality of life. Other studies suggest that patients affected by a condition but having more severe discomfort or depression give lower condition-specific utilities30,31; these factors could also explain our findings.
Our study has certain limitations. Most of our subjects were recruited from a sexually transmitted disease clinic in one US city, which may limit the generalizablility of our findings. Only a small number of study subjects had experienced the health states of ectopic pregnancy, chronic pelvic pain, or infertility, so that we may not have truly captured a patient’s perspective for those health states. A 10-year time horizon was used for TTO valuation of both short- and long-term health states, potentially leading to different valuations for the short-term states (PID treatment and ectopic pregnancy) if more complex, chained methods of valuation had been performed.32 Appropriate use of these valuations in future cost-effectiveness analyses requires appreciating the differences between short-term and long-term health states. Finally, few subjects “repaired” their utilities when they rated a health state as worse than the ambulatory treatment health state, which we considered, a priori, the least unfavorable health state. However, when subjects were asked why they rated ambulatory treatment lower, they generally gave logical reasons. The most common situation was rating hospital treatment of PID as better than ambulatory treatment, with some women clearly preferring hospital care.
We conclude that PID has a substantial impact on health utility. Women who have experienced PID have lower valuations for PID complication health states, although those differences only reached statistical significance on the VAS. Nevertheless, differences in utilities could have some impact on cost-effectiveness analyses of PID treatment or prevention if the analysis is conducted from the societal perspective as opposed to the patient perspective.
Appendix 1: Demographic, Childbearing, and PID Symptom-Related Information
Please enter your age:
Please select your marital status: (Single, Married, Divorced, Separated, Widowed)
Please select your ethnicity: (Hispanic or Latino, Not Hispanic or Latino, I do not wish to answer)
Please select your race: (American Indian/Alaska Native, Asian, Native Hawaiian or Other Pacific Islander, black or black, white, More than one race, I do not wish to answer)
Please indicate your educational background: (Completed Primary Education, Some High School, High School Diploma, Some College, Associate Degree, Bachelor’s Degree, Master’s Degree, Doctorate or Terminal Degree, I do not wish to answer)
Are you employed outside your home? (No, Yes)
Have you ever had PID? (No, Yes)
Are you trying to get pregnant but not able to? (No, Yes)
Do you have children? (No; Yes, 1 child; Yes, 2 or more children)
Do you want to have children (or more children) in the future? (No, Yes)
Do you have chronic pain in your lower abdomen and pelvis? (No, Yes)
Have you had an ectopic (or tubal) pregnancy? (No, Yes) Cited Here...
Appendix 2: PID Health State Scenarios
PID Outpatient Treatment Scenario
Imagine a 25-year-old woman with PID that does not require a hospital stay for treatment. She will take antibiotic pills for 14 days to treat it.
She will have pain for about 7 days, with the pain mainly in the lower abdomen. It will interfere with daily activity, work, sleep, and family relations. Having sex is usually painful, and fever, nausea, and vaginal discharge are often part of the illness. She will be able to eat and drink pretty much as usual.
She has a very small chance of developing complications that could require a hospital stay and possibly an operation. Long-term problems with pain, difficulty with becoming pregnant, or with tubal pregnancy could occur, even if she is treated.
However, she will probably return to her usual health once the illness goes away, but will have some chance of getting PID again in the future.
PID Inpatient Treatment Scenario
Imagine a 25-year-old woman with PID that requires a hospital stay for treatment. She will be in the hospital for 2 or 3 days and will receive antibiotics by vein (through an IV) during this time. After she goes home, she will take antibiotic pills for 11 or 12 more days.
She will have pain for about 7 days, with more pain than with illness not requiring a hospital stay. The pain is mainly in the lower abdomen and severe enough to interfere with daily activity, work, sleep, and family relations. Having sex is usually painful, and fever, nausea, and vaginal discharge are often part of the illness. She will soon be able to eat and drink pretty much as usual.
She has a small chance of developing complications that could require a longer hospital stay and possibly an operation. Long-term problems with pain, difficulty with becoming pregnant, or with tubal pregnancy could occur, even if she is treated. Complications and other problems happen more often when the illness requires a hospital stay.
However, she will probably return to their her usual health once the illness goes away, but will have some chance of getting PID again in the future.
Ectopic Pregnancy Scenario
Imagine a 25-year-old woman who has an ectopic (or tubal) pregnancy. This is a pregnancy that is not in the womb, but is abnormally located in one of the tubes leading to the womb.
She can have pain in the abdomen and will need to be treated to remove the abnormal pregnancy. Some women will need to have a minor operation that lasts for about an hour and involves small cuts (or incisions) to remove the pregnancy. They can go home right after this operation. Other women can get 1 or 2 shots of a medicine in the arm that removes the abnormal pregnancy over the next several weeks. There is a very small chance that the tube could burst, requiring an emergency operation to remove both the tube and the abnormal pregnancy.
After the abnormal pregnancy is removed, she will soon be without pain and will be able to go back to her usual activities in the next few weeks.
Because of the abnormal pregnancy, there is a chance that getting pregnant in the future will be harder. She will have a greater chance of having this kind of pregnancy again in the future.
Chronic Pelvic Pain Scenario
Imagine a 25-year-old woman with continuing pain in her lower abdomen and pelvic area.
Because of this pain, she is limited a little in moderate activities, such as moving a table or pushing a vacuum cleaner, and limited a lot in climbing several flights of stairs.
She often feels downhearted and blue. She wishes she could feel more calm and peaceful, and have more energy.
The pain makes her do less than she wants and interferes with work (both outside the home and housework) and social activities (visiting with friends, relatives, etc.). The pain makes having sex difficult.
The pain may slowly go away as time goes on but could also stay the same.
Imagine a 25-year-old woman who has been trying to get pregnant for 1 year but has not been able to.
She has no problems with pain or with doing things she wants to do, but feels stressed because she wants to have children soon. She wishes she could feel more calm and peaceful.
Because she has not been able to get pregnant, she feels less satisfied with her partner, her sex life, and her overall quality of life. Cited Here...
1. Centers for Disease Control and Prevention. Sexually transmitted diseases treatment guidelines 2006. MMWR Morb Mortal Wkly Rep 2006; 55(RR11):1–94.
2. Beigi RH, Wiesenfeld HC. Pelvic inflammatory disease: New diagnostic criteria and treatment. Obstet Gynecol North Am 2003; 30:777–793.
3. Rein DB, Kassler WJ, Irwin KL, et al. Direct medical cost of pelvic inflammatory disease and its sequelae: decreasing, but still substantial. Obstet Gynecol 2000; 95:397–402.
4. Gold MR, Siegel JE, Russell LB, et al., eds. Cost-Effectiveness in Health and Medicine. New York: Oxford University Press, 1996.
5. Gold MR, Franks P, McCoy KI, et al. Toward consistency in cost-utility analyses: Using national measures to create condition-specific values. Med Care 1998; 36:778–792.
6. Torrance GW. Utility approach to measuring health-related quality of life. J Chronic Dis 1987; 40:593–603.
7. Tsevat J. What do utilities measure? Med Care 2000; 38:II160–II164.
8. Hu D, Hook EW III, Goldie SJ. Screening for Chlamydia trachomatis in women 15 to 29 years of age: A cost-effectiveness analysis. Ann Intern Med 2004; 141:501–513.
9. Aledort JE, Hook EW III, Weinstein MC, et al. The cost effectiveness of gonorrhea screening in urban emergency departments. Sex Transm Dis 2005; 32:425–436.
10. Appendix 2: Chlamydia. In: Stratton KR, Durch J, Lawrence RS, eds. Vaccines for the 21st Century: A Tool for Decisionmaking. Washington, DC: National Academy Press, 2000.
11. Appendix 17: Neisseria gonnorrhea. In: Stratton KR, Durch J, Lawrence RS, eds. Vaccines for the 21st Century: A Tool for Decisionmaking. Washington, DC: National Academy Press, 2000.
12. Saha S, Hoerger TJ, Pignone MP, et al. The art and science of incorporating cost effectiveness into evidence-based recommendations for clinical preventive services. Am J Prev Med 2001; 20:36–43.
13. Ness RB, Soper DE, Peipert J, et al. Design of the PID Evaluation and Clinical Health (PEACH) Study. Control Clin Trials 1998; 19:499–514.
14. Ness RB, Soper DE, Holley RL, et al. Effectiveness of inpatient and outpatient treatment strategies for women with pelvic inflammatory disease: Results from the Pelvic Inflammatory Disease Evaluation and Clinical Health (PEACH) Randomized Trial. Am J Obstet Gynecol 2002; 186:929–937.
15. Lenert LA, Sturley A, Watson ME. iMPACT3: Internet-based development and administration of utility elicitation protocols. Med Decis Making 2002; 22:464–474.
16. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: Construction of scales and preliminary tests of reliability and validity. Med Care 1996; 34:220–233.
17. Lenert LA, Sturley A, Rupnow M. Toward improved methods for measurement of utility: Automated repair of errors in elicitations. Med Decis Making 2003; 23:67–75.
18. Walters SJ, Brazier JE. Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D. Qual Life Res 2005; 14:1523–1532.
19. Torrance GW, Boyle MH, Horwood SP. Application of multi-attribute utility theory to measure social preferences for health states. Oper Res 1982; 30:1043–1069.
20. Nord E. The validity of a visual analogue scale in determining social utility weights for health states. Int J Health Plann Manage 1991; 6:234–242.
21. Green C, Brazier J, Deverill M. Valuing health-related quality of life. A review of health state valuation techniques. Pharmacoeconomics 2000; 17:151–165.
22. Torrance GW, Feeny D, Furlong W. Visual analog scales: do they have a role in the measurement of preferences for health states? Med Decis Making 2001; 21:329–334.
23. Froberg DG, Kane RL. Methodology for measuring health-state preferences. II. Scaling methods. J Clin Epidemiol 1989; 42:459–471.
24. Songer TJ, Lave JR, Kamlet MS, et al. Preferences for fertility in women with pelvic inflammatory disease. Fertil Steril 2004; 81:1344–1350.
25. Tengs TO, Wallace A. One thousand health-related quality-of-life estimates. Med Care 2000; 38:583–637.
26. Fryback DG, Dasbach EJ, Klein R, et al. The Beaver Dam Health Outcomes Study: Initial catalog of health-state quality factors. Med Decis Making 1993; 13:89–102.
27. Sullivan PW, Ghushchyan V. Preference-based EQ-5D index scores for chronic conditions in the United States. Med Decis Making 2006; 26:410–420.
28. Ubel PA, Loewenstein G, Hershey J, et al. Do nonpatients underestimate the quality of life associated with chronic health conditions because of a focusing illusion? Med Decis Making 2001; 21:190–199.
29. Tsevat J, Sherman SN, McElwee JA, et al. The will to live among HIV-infected patients. Ann Intern Med 1999; 131:194–198.
30. Melsop KA, Boothroyd DB, Hlatky MA. Quality of life and time trade-off utility measures in patients with coronary artery disease. Am Heart J 2003; 145:36–41.
31. Wells KB, Sherbourne CD. Functioning and utility for current health of patients with depression or chronic medical conditions in managed, primary care practices. Arch Gen Psychiatry 1999; 56:897–904.
32. Jansen SJ, Stiggelbout AM, Wakker PP, et al. Patients’ utilities for cancer treatments: A study of the chained procedure for the standard gamble and time tradeoff. Med Decis Making 1998; 18:391–399.