Shih, Ting-Yu BS*; Gaydos, Charlotte A. MS, MPH, DrPH*†; Rothman, Richard E. MD, PhD*†; Hsieh, Yu-Hsiang PhD*
In the United States, pelvic inflammatory disease (PID) affects approximately 1 million women every year.1 PID includes a broad spectrum of upper female genital tract disorders, and its diagnosis and treatment are often complicated.2 Thus, the Centers for Disease Control and Prevention (CDC) provided treatment guidelines for PID in 1991,3 and updated the guidelines 4 times since then.2,4–6 However, the guidelines are complicated; the CDC recommended 4 to 6 different regimens, and each regimen included different combinations of medications. Consequently, substantial divergence from the guidelines occur among physicians including those who practice emergency medicine, family practice, general practice, internal medicine, obstetrics and gynecology, and pediatrics.7 Improper treatment of PID may cause many serious long-term sequelae such as infertility8 and ectopic pregnancy.9 Healthy People Initiative, which was led by government agencies including the CDC and the National Institutes of Health to focus public and private efforts to address the most significant preventable threats to health, aimed to have at least 90% of the primary care providers to manage sexually transmitted diseases (STDs) correctly by the year of 2000,10 and the same goal was targeted for 2010.11
PID is commonly diagnosed in emergency departments (EDs) with approximately 340,000 to 410,000 cases per year.12,13 Among these visits, approximately 70% to 90% of the patients were discharged from EDs.14,15 These patients would be prescribed 2 weeks of outpatient medications under CDC-recommended treatment guidelines. Nevertheless, studies have demonstrated that the ED provider adherence rates to CDC treatment guidelines for PID for discharged patients were only 32% to 35%.14,15 However, these 2 studies are limited in that they either involved a single-hospital with 10 months of data from the early 2000s14 or were drawn from a national representative study of adolescents from 1992 to 1998 with significant temporal biases where adherence was determined by applying treatment guidelines that were published later in 1998.15 Little is known about more recent ED provider adherence rates with the CDC guidelines for PID among both adults and adolescents on the national level since 1998.
In this study, we analyzed the National Hospital Ambulatory Medical Care Survey (NHAMCS) database sampled from all US EDs from 1999 to 2006 to (1) estimate national rates of provider adherence to the CDC treatment guidelines for STDs among patient visits with a diagnosis of acute PID in US ED settings, and (2) determine the factors associated with provider nonadherence.
We conducted this cross-sectional study evaluating rates and factors associated with provider adherence to CDC treatments guidelines for PID among US ED visits, using the ED component of the NHAMCS database from 1999 to 2006. The NHAMCS is a national probability sample survey of nonfederal, general, and short stay hospitals conducted by the Division of Health Care Statistics of the National Center for Health Statistics, Centers for Disease Control and Prevention.16 The survey has a 4-staged sample design: geographic primary sampling units, hospitals with EDs within primary sampling units, emergency service areas within EDs, and patient visits within emergency service areas. Patient visit data included patient demographics, expected source of payment, provider type including physicians, nurse practitioners, and physician assistants, up to 6 medications from 1999 to 2002 and up to 8 medications from 2003 to 2006, which could be given during the visit in the ED and/or prescribed upon discharge from the ED, up to 3 ED diagnoses, and patient disposition as recorded by trained hospital staff or Census Bureau field representatives, using a systematic random sample over a randomly assigned 4-week reporting period.16 Data consistency was routinely verified. Internal NHAMCS check on data entry and coding found very low error rates.17 The ED component of the NHAMCS data are used to statistically describe the visits that use hospital ED services, and are used to present and understand the changes that occur in medical care requirements and practices.16 The study was granted exempt from review status by Johns Hopkins University School of Medicine Institutional Review Board, due to the public accessibility and the deidentified characteristics of the database.
Only female patients older than 12 years old were included in the study. We defined a PID visit as having a diagnosis of acute PID as coded by International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 098.10, 098.16, 098.17, 098.86, 614.0, 614.2, 614.3, 614.5, 614.8, 614.9, 615.0, 615.9.18 During the 8 year study period, data from 118,740 ED visits were collected, among which 487 visits were diagnosed with PID.
The 1998, 2002, and 2006 CDC treatment guidelines for PID were summarized in Table 1. The 1998 guidelines were in effect until May 2002 when the 2002 guidelines were announced while the 2002 guidelines were in effect until August 2006 when the 2006 guidelines were announced. New treatment guidelines are expected to be released in 2010. Specific antibiotics ordered or provided at each visit were identified and compared to the treatment guidelines (1998 guidelines for visits during January 1999–May 2002; 2002 guidelines for visits during June 2002–August 2006; 2006 guidelines for visits during September 2006–December 2006) to determine rates of provider adherence to CDC treatment guidelines. During the whole study period, regardless of which treatment guidelines, we further categorized provider adherence into 3 categories: (1) complete provider adherence, if the patient visit included prescriptions with any of the CDC-recommend regimens; (2) partial provider adherence, if the patient visit included only part of the CDC-recommended regimens; (3) complete nonadherence, if the patient visit did not include any of the medications included in the CDC-recommended regimens. For simplicity, we defined partial provider adherence and complete nonadherence collectively as nonadherence as compared to complete provider adherence. The provider type was defined operationally as 1 of 3 different categories of physicians: (1) ED attending/staff physician, including staff physician category from 1999 to 2004 and attending physician category from 2005 to 2006; (2) ED resident/intern, including resident/intern category from 1999 to 2004 and ED resident/intern category from 2005 to 2006; (3) non-ED physician, who was not an ED attending/staff, resident, or intern physician; that is, a specialist consultant to the ED, including other physician category from 1999 to 2004 and on call attending physician/fellow category from 2005 to 2006. Provider types were not mutually exclusive; that is, a visit could be seen by multiple provider types.
We merged the NHAMCS data from year 1999 to 2006 for data analysis. A sample weight which considered selection probability, nonresponse adjustment, and ratio adjustment for different total sample size each year, was assigned for each patient visit.16 Although the sampling fraction was relatively small, the weighted numbers calculated by the method suggested by the CDC, which accounts for the 4-stage probability sampling scheme, represented an unbiased national estimate of the US ED population, if it was based on more than 30 cases in the sample data or if the relative standard error was greater than 30%.16 Adjusted sample weights, strata, and primary sampling unit design variables provided by NHAMCS were used in all analyses by using SAS 9.1 SURVEYFREQ and SURVEYLOGISTIC procedures (SAS Institute Inc., Cary, NC). Results were reported as weighted frequencies, percentages, and 95% confidence intervals (CI) for individual characteristics of interest. Age, gender, race, and ethnicity specific census data were obtained from the US Census Bureau in order to estimate visit rates.19 Only those visits that resulted in patient discharge were included in bivariate analyses, which were performed to determine the adherence to CDC STD treatment guidelines during PID ED visits by sociodemographics, ED characteristics, ED visit characteristics, and provider types. Bivariate analysis and estimation of 95% CI were not conducted if the sample size of interest was less than 30, as the estimate is considered unreliable under NHAMCS analysis recommendations.16 Multivariate logistic regression analysis was performed to determine the factors associated with ED medical care provider's nonadherence to CDC STD treatment guidelines. P < 0.05 was considered statistically significant. The data for visits after the announcement of the 2006 guidelines were combined with those for visits after the announcement of the 2002 guidelines for analysis due to small sample size. All percentages presented were weighted percentages unless indicated.
During the 8 survey years, 1,755,000 (95% CI, 1,533,000–1,977,000) ED visits resulted in an ED diagnosis of acute PID among 399,168,000 (95% CI, 371,808,000–426,528,000) ED visits of female patients who were older than 12 years. There were approximately 219,000 PID-related visits per year. The estimated numbers of PID visits by demographics and disposition are presented in Table 2. The overall PID visit rate of the study population was 180.8 in 100,000 visits. Women aged 20 to 29 years and black had the highest PID visit rate (Table 2).
Among the 1,755,000 visits with a diagnosis of acute PID, 1,605,000 visits ([95% CI, 1,403,000–1,808,000]; weighted percentage, 91.5% [95% CI, 88.4%–94.5%]) were discharged from the ED. The ED provider adherence to CDC treatment guidelines for PID among these visits was investigated. Among them, 492,000 ([95% CI, 381,000–603,000]; weighted percentage, 30.7% [95% CI, 24.5%–36.8%]) of visits had full provider adherence, 758,000 ([95% CI, 605,000–911,000]; weighted percentage, 47.2% [95% CI, 40.8%– 53.7%]) of visits had partial provider adherence, and 355,000 ([95% CI, 260,000–450,000]; weighted percentage, 22.1% [95% CI, 16.6%–27.6%]) visits had complete nonadherence. Bivariate analysis showed that providers for visits to EDs located in the US Northwest, Midwest and South geographic regions, for visits before the announcement of the 2002 guidelines, and for those that were not seen by non-ED physician (i.e., not seen by specialist consultants) were significantly more likely to have nonadherence to CDC treatment guidelines (Table 3). In multivariate analysis adjusted for other covariates, providers for ED visits before the announcement of 2002 CDC treatment guidelines, EDs located in the US Northeast, Midwest, and South geographic regions, and not seen by non-ED physicians were significantly more likely to have nonadherence to the CDC PID treatment guidelines than their counterparts (Table 4).
We conducted further analyses to study the pattern of the regimens prescribed to PID patients (Table 5). Doxycycline was the drug that was missing most frequently from the recommended regimens among visits with partial provider adherence (66.5%; 95% CI, 59.1%–73.9%). Azithromycin, which was not a CDC-recommended medication for PID in the 1998 and 2002 guidelines, was the most prescribed drug during those visits with nonadherence (32.6%; 95% CI, 25.0%–40.1%). Additionally, prescription patterns for visits with complete nonadherence were summarized in Table 6. Among them, 29.7% were given antibiotics that were not CDC-recommended medications, 33.0% were prescribed pain relievers without any antibiotics among which 79.4% only had pain reliever prescriptions, while 37.3% were prescribed neither antibiotics nor pain relievers. Of note, 57.3% of the visits with complete nonadherence were given PID as the primary ED diagnosis.
We reported a 69.5% nonadherence, that is, 30.5% complete provider adherence rate, to the CDC PID treatment guidelines among the ED visits with a diagnosis of PID, similar to findings from a single-hospital study that reviewed the charts of ED visits with a diagnosis of PID from 2000 to 2001,14 although the authors in that study used a single ICD-9 code (614.9) only for their definition of PID. Notably, we reported even a lower adherence rate among visits (24.1%) by adolescents (less than 20 years of age). Our findings differ from that reported by Beckmann et al., which showed a 35% provider adherence rate for adolescents, using the 1992–1998 NHAMCS database.15 That study may have been somewhat temporally skewed, however, because they referenced subsequent CDC treatment guidelines as the standard of treatment for the study's ED practices from prior years. In addition, they did not report how they defined PID by using ICD-9 codes. Therefore, how they captured PID cases from NHAMCS database remains unclear.
We reported a lower annual average number of PID visits than in previous literature. It might be because our study focused on acute PID visits while the literature included both acute and chronic PID diagnoses.12,13 Moreover, our study period was newer than one of the studies,12 and the other study included ICD-9 code 616, which is not “traditionally” classified as PID.13
One of the reasons for the low adherence rates might be because ED providers simply were not aware that they were not following current CDC guidelines as suggested by Hessol's study on the management of PID by primary care physicians.7 Timely communication and education between the CDC and the national specialty societies (here the American College of Emergency Physicians) with subsequent transmission of that information to individual institutions may help to circumvent potential knowledge gaps.20 Utilization of information technology, such as computer-assisted decision support systems, or electronic medical records (with monitoring and provision of feedback to providers) may provide a method that could be leveraged to educate and encourage provider adherence to guidelines.21–23 An alternative explanation could be that without insurance patients might not be able to afford the CDC-recommended antibiotics, so ED physicians did not prescribe these medications. However, our data showed no difference in adherence rate (P > 0.05) between visits with or without insurance. Thus, the possibility that some CDC-recommended antibiotics were not prescribed due to lack of insurance coverage was less likely.
Our results demonstrated that ED visits after the announcement of 2002 CDC guidelines were more likely to have complete provider adherence to the CDC guideline than those before. This implied that complete ED provider adherence rates with the CDC guidelines might be showing signs of improvement. To explain the improvement and to investigate whether this represents a true trend will require future study. Interestingly, during our study period, ED visits that were seen by non-ED physicians (specialists) had higher complete provider adherence rate. A possible explanation might be that the observed improved adherence may be attributable to care rendered by consultants (likely principally obstetrics-gynecology specialists). Such specialists could help in increasing ED provider adherence to the PID treatment guidelines in the future, by participating in educational programs for ED providers.
This study also reported that ED providers in the US West geographical region were more likely to demonstrate adherence to the CDC guidelines than other US regions. This implies some regional variations, which might be influenced by state policies or hospital systems that encouraged adherence to treatment guidelines, and/or by practice behaviors that were more prone to adherence to guidelines by some clusters of individual providers, who received continued medical education training focusing on CDC treatment guidelines. One possibility is that in the West,24 managed care organizations are most common. Such organizations can influence the likelihood that patients receive treatments recommended by evidence-based guidelines by deciding what treatments and services will be reimbursed.25 However, the NHAMCS database limits our ability to investigate this issue further. Future studies are needed to determine why the West was more likely to have provider adherence. The experience in this region could potentially help enhance adherence rates in other regions.
We discovered that azithromycin was the leading antibiotic prescribed to patients without full adherence to the CDC treatment guidelines. Our finding that 32.7% of the visits without full adherence to the CDC guidelines were prescribed azithromycin was virtually the same as the finding by Kane et al.14 One possible explanation could be that the widespread use of azithromycin for treatments of chlamydial infections.26 Physicians might be so accustomed to using azithromycin for uncomplicated STDs that they might not recognize the need for other regimens for more complicated infections such as PID. While technically not adherent to guidelines, prescribing azithromycin did comply with the intention for the treatment of chlamydia. As a monotherapy it was problematic since other possible microbials causing PID were not covered. If it was used as part of a combination therapy which also provided coverage against other PID-associated pathogens, it might be more appropriate. An alternative explanation might be that physicians believed that azithromycin therapy was as good as or better than doxycycline therapy. One randomized clinical trial did show that 7-day azithromycin monotherapy and 12-day azithromycin plus metronidazole combination therapy had higher clinical or microbiologic cure rates than 14-day or 21-day standardized multidrug regimens including doxycycline.27 However, this study was designed to assess inpatient, not outpatient treatment regimens.28 Furthermore, the validity and generalizability of the result decrease because 78% of the study population dropped out at the final follow-up.26,29 A third explanation might be that providers perceived that patient adherence to treatment might be higher with azithromycin therapy which required shorter courses and fewer doses.26,30 Notably, it was stated in the 2006 guidelines that azithromycin, with metronidazole cotherapy, had demonstrated high short-term clinical cure rates for PIDs as an alternative parenteral regimen or oral regimen A.6
Doxycycline was not given in the majority of the visits with partial provider adherent to the CDC guidelines. No previous literature reported why doxycycline was often missing from treatments for PID patients. One explanation is that perhaps the missing prescription resulted from treatment failure or side effects caused by doxycycline prescribed in previous medical visits,31 so that ED providers considered other medications for the present visit. Since our study was a cross-sectional study, we did not have patients' medical history to determine whether resistance or side effects were the reason for not prescribing doxycycline during these visits. Another plausible explanation drawn from empirical observations by one of the coauthors (R.R.) is that ED providers often give empirical azithromycin in the ED based on the belief that patients will not adhere to 2 weeks of doxycycline, and then either forget or choose not to give the recommended 2-week course of doxycycline at the time of discharge.
The validity of ED diagnosis of acute PID could affect our estimate of provider adherence rates. The diagnosis of acute PID is difficult due to the wide spectrum of signs and symptoms. A more accurate diagnosis of PID as well as bacteriologic diagnosis requires a more invasive approach, e.g., laparoscopy, which is not practical in the ED settings; therefore, the main diagnostic method is based on clinical findings.2 Consequently, an ED visit might be clinically misclassified as PID when it was actually not, or misclassified as non-PID but it actually was. The former might be overtreated and the latter might receive inadequate treatment, which could lead us to overestimation or underestimation of provider adherence rate. However, if a patient is diagnosed with PID, the treatment should be given according to the CDC guidelines. A scenario study where ED physicians diagnose and provide treatments based on the scenarios presented can be used to assess whether they know about the CDC guidelines and why they do not follow the guidelines if they do know about the guidelines. Furthermore, clinical trials can be used to evaluate whether information technology helps promote adherence to treatment guidelines.
Our study is subject to some limitations. First, the NHAMCS database did not provide information regarding actual dosage of prescribed antibiotics or course of treatment given to patients. A treatment should be counted as nonadherence to the CDC guidelines if the correct medications were given to patients at wrong doses or courses, which had been reported in Kane's study (22%).14 This would lead to underestimation of the nonadherence rate to the CDC treatment guidelines in EDs. Another limitation of our study is that other clinical information including patients' allergies, the severity and history of PID, and previous treatment outcomes, which could influence ED physicians' decisions for treatments, was not available in the NHAMCS database. Therefore, the extent that our measurement of provider adherence to the CDC guidelines was affected by ED physicians' decisions based on the above information is unknown. Moreover, coding of the diagnoses could be driven partly by billing, where physicians still coded a patient with a low suspicion of PID as PID, and subsequently they correctly decided that antibiotics were not indicated. This would influence our measurement of provider adherence as well. Lastly, up to 6 or 8 drugs per visit depending on the survey year were recorded in the NHAMCS database, so some of the prescribed PID antibiotics might not be recorded in the survey if a visit was prescribed with more than 8 drugs, implying that some of the visits with partial provider adherence might have had full provider adherence. However, this should have little impact on our measurement of provider adherence since most of the visits diagnosed with PID were prescribed with fewer than 6 drugs.
This study showed overall low rates of ED provider adherence to the CDC treatment guidelines for PID in US EDs. Frequently, treatment regimens were not adherent with the particular antibiotics recommended in the guidelines completely, and sometimes antibiotics were not even prescribed at all. The gap between what should be prescribed and what was actually prescribed in EDs, raises important quality of care and public health concerns. Efforts should be directed to develop strategies that can enhance the implementation and consistent adherence to CDC recommendations for PID treatment guidelines in our nation's EDs.
2.Sexually transmitted diseases treatment guidelines 2002. Centers for Disease Control and Prevention. MMWR Recomm Rep 2002; 51(RR-6):1–78.
3.Pelvic inflammatory disease: Guidelines for prevention and management. MMWR Recomm Rep 1991; 40(RR-5):1–25.
4.1993 sexually transmitted diseases treatment guidelines. Centers for Disease Control and Prevention. MMWR Recomm Rep 1993; 42(RR-14):1–102.
5.1998 guidelines for treatment of sexually transmitted diseases. Centers for Disease Control and Prevention. MMWR Recomm Rep 1998; 47(RR-1):1–111.
6.Sexually transmitted diseases treatment guidelines 2006. Centers for Disease Control and Prevention. MMWR Recomm Rep 2006; 55(RR-11):1–94.
7.Hessol NA, Priddy FH, Bolan G, et al. Management of pelvic inflammatory disease by primary care physicians. A comparison with Centers for Disease Control and Prevention guidelines. Sex Transm Dis 1996; 23:157–163.
8.Hillis SD, Joesoef R, Marchbanks PA, et al. Delayed care of pelvic inflammatory disease as a risk factor for impaired fertility. Am J Obstet Gynecol 1993; 168:1503–1509.
9.Marchbanks PA, Annegers JF, Coulam CB, et al. Risk factors for ectopic pregnancy. A population-based study. JAMA 1988; 259:1823–1827.
10.U.S. Department of Health and Human Services. Healthy people 2000: National health promotion and disease prevention objectives. Full report with commentary. Department of Health and Human Services publication PHS 91–60212. Washington, DC: U.S. Department of Health and Human Services, 1991.
12.Curtis KM, Hilliss SD, Kieke BA, et al. Visits to emergency departments for gynecologic disorders in the United States, 1992–1994. Obstet Gynecol 1998; 91:1007–1012.
13.Schappert SM, Rechtsteiner EA. Ambulatory medical care utilization estimates for 2006. Natl Health Stat Report 2008; 8:1–29.
14.Kane BG, Degutis LC, Sayward HK, et al. Compliance with the Centers for Disease Control and Prevention recommendations for the diagnosis and treatment of sexually transmitted diseases. Acad Emerg Med 2004; 11:371–377.
15.Beckmann KR, Melzer-Lange MD, Gorelick MH. Emergency department management of sexually transmitted infections in US adolescents: Results from the National Hospital Ambulatory Medical Care Survey. Ann Emerg Med 2004; 43:333–338.
16.National Center for Health Statistics, Centers for Disease Control and Prevention. Ambulatory Health Care Data: National Hospital Ambulatory Medical Care Survey. Available at: http://www.cdc.gov/nchs/ahcd/about_ahcd.htm
. Accessed August 5, 2009.
17.Nawar EW, Niska RW, Xu J. National Hospital Ambulatory Medical Care Survey: 2005 emergency department summary. Adv Data 2007; 386:1–32.
18.Sutton MY, Sternberg M, Zaidi A, et al. Trends in pelvic inflammatory disease hospital discharges and ambulatory visits, United States, 1985–2001. Sex Transm Dis 2005; 32:778–784.
20.Hsieh YH, Chen KF, Gaydos CA, et al. Antiviral prescriptions to U.S. ambulatory care visits with a diagnosis of influenza before and after high level of adamantane resistance 2005–06 season. PLoS One 2010; 5:e8945.
21.Asaro PV, Sheldahl AL, Char DM. Embedded guideline information without patient specificity in a commercial emergency department computerized order-entry system. Acad Emerg Med 2006; 13:452–458.
22.Ozdas A, Speroff T, Waitman LR, et al. Integrating “best of care” protocols into clinicians' workflow via care provider order entry: Impact on quality-of-care indicators for acute myocardial infarction. J Am Med Inform Assoc 2006; 13:188–196.
23.Grol R. Improving the quality of medical care: Building bridges among professional pride, payer profit, and patient satisfaction. JAMA 2001; 286:2578–2585.
24.Lagoe R, Aspling DL, Westert GP. Current and future developments in managed care in the United States and implications for Europe. Health Res Policy Syst 2005; 3:4.
25.Curry SJ. Organizational interventions to encourage guideline implementation. Chest 2000; 118(suppl 2):40S–46S.
26.Haggerty CL, Ness RB. Diagnosis and treatment of pelvic inflammatory disease. Womens Health (Lond Engl) 2008; 4:383–397.
27.Bevan CD, Ridgway GL, Rothermel CD. Efficacy and safety of azithromycin as monotherapy or combined with metronidazole compared with two standard multidrug regimens for the treatment of acute pelvic inflammatory disease. J Int Med Res 2003; 31:45–54.
28.Savaris RF, Teixeira LM, Torres TG, et al. Comparing ceftriaxone plus azithromycin or doxycycline for pelvic inflammatory disease: A randomized controlled trial. Obstet Gynecol 2007; 110:53–60.
29.Haggerty CL, Ness RB. Newest approaches to treatment of pelvic inflammatory disease: A review of recent randomized clinical trials. Clin Infect Dis 2007; 44:953–960.
30.Walling AD. Tips from other journals: Azithromycin vs. doxycycline for the treatment of PID. Am Fam Physician 2008; 77:834–837.
31.Walker CK, Workowski KA, Washington AE, et al. Anaerobes in pelvic inflammatory disease: Implications for the Centers for Disease Control and Prevention's guidelines for treatment of sexually transmitted diseases. Clin Infect Dis 1999; 28(suppl 1):S29–S36.