Pregnancy Reasonably Excluded Guide (PREG) Evaluation of Pregnancy Status Before Contraceptive Procedures: Improved Availability of Same-Day Insertion.

We collected data from 1012 visits among 981 women scheduled for IUC or implant placement during the study period. Overall, the mean age of women screened with PREG was 30.6 years; at 1012 visits, 880 patients (87.0%) identified as white, 533 (52.7%) reported gravidity of 0, and 557 (55.0%) reported parity of 0 ( Table 1 ). Most of the contraceptive procedures (812 of 1012 [80.2%]) were IUC placement; the other 19.8% of procedures (200) were placement of the contraceptive implant etonogestrel, 68 mg ( Table 1 ). Twenty women (1.9%) had contraceptive procedures after bilateral tubal ligation or menopause for off-label use. 11 Table 1 Demographic Characteristics, Pregnancy History, and Outcomes for Primary Care Gynecology Clinic Visits a , b Variable Value (N=1012 visits) Patient age at visit (y) 30.6 (8.6) Race  White 878 (86.8)  Black/African American 33 (3.3)  American Indian/Alaskan Native 2 (0.2)  Asian/Pacific Islander 47 (4.6)  Other 42 (4.1)  Not disclosed/unknown 10 (1.0) Hispanic or Latina  Yes 36 (3.6)  No 944 (93.3)  Not disclosed/unknown 32 (3.2) Gravidity 0 (0-2) Parity 0 (0-2) Type of contraception  Etonogestrel implant, 68 mg 200 (19.8)  IUC 812 (80.2)  Levonorgestrel-releasing intrauterine system, 52 mg 711 (70.3)  Intrauterine copper contraceptive 49 (4.8)  Levonorgestrel-releasing intrauterine system, 13.5 mg 51 (5.0)  Levonorgestrel-releasing intrauterine system, 19.5 mg 1 (0.1) Point-of-care pregnancy test 244 (24.1)  hCG positive 1 (0.4) (0.1% of cohort) Rescheduled  No 974 (96.2)  Not recorded 2 (0.2)  Yes 36 (3.6) a hCG = human chorionic gonadotropin; IUC = intrauterine contraceptive. b Data are presented as mean ± SD, No. (percentage) of visits, or median (interquartile range). Demographic Characteristics, Pregnancy History, and Outcomes for Primary Care Gynecology Clinic Visits a , b hCG = human chorionic gonadotropin; IUC = intrauterine contraceptive. Data are presented as mean ± SD, No. (percentage) of visits, or median (interquartile range). At 1012 visits after PREG implementation, 974 women (96.2%) were eligible for their scheduled contraceptive procedure ( Table 1 ). Only 36 women (3.6%) had their appointment rescheduled because of uncertain pregnancy status, and 2 women (0.2%) did not have this information recorded in the medical record. If traditional criteria, current menses, or presence of a current IUC or implant had been used to exclude pregnancy, only 594 (58.7%) would have qualified for a same-day procedure. No contraceptive procedures occurred in pregnant women. No failures to detect pregnancy were identified after the procedures based on the absence of reports to the PCG clinic, obstetric health care professionals, and quality offices. Before PREG implementation, 98 of 185 women (53.0%) in a historical cohort had hCG measurement. Use of PREG decreased the rate of hCG testing to 24.1% (244 tests per 1012 visits) with point-of-care urine testing and an additional 3.5% (35 tests per 1012 visits) with laboratory testing ordered in advance. Of the 244 point-of-care hCG tests, 1 was positive in the context of a group C response, and the contraceptive procedure was postponed. Measurement of hCG after PREG implementation was grouped by age and procedure type. Women aged 18 to less than 26 years had the highest rate of office hCG tests (114 of 403 [28.3%]) compared with age 26 to less than 40 years (101 of 433 [23.3%]), age 40 to less than 46 years (17 of 125 [13.6%]), and age 46 to less than 51 years (12 of 51 [23.5%]). Women scheduled for implant placement had a higher rate of hCG tests (55 of 200 [27.5%]) than those scheduled for IUC placement (189 of 812 [23.3%]). The rate of intended IUC placement increased with increasing age ( Figure 3 ). Figure 3 Percentage of intrauterine contraceptive users by age. Percentage of intrauterine contraceptive users by age. PREG responses by category are summarized in Table 2 . Among the 1012 visits, 538 patients (57.6%) had group B responses only, 145 (14.3%) had group A only, 144 (14.2%) had responses in both groups A and B, and 105 (10.4%) had group C responses. “I think I may be pregnant” or “I would like a pregnancy test” (a subset of group C) was the response for 4 women (0.4%). On the basis of PREG responses alone without APP review, 121 women (12.0%) would have required pregnancy testing. Provider review and discussion of the PREG led to a 24.1% rate of urine hCG testing in the clinic (244 of 1012 visits), with an additional 3.5% (35 patients) having laboratory pregnancy testing ordered before the procedure visit by the referring physician. Indications for testing after provider review of PREG included identification of an expired LARC, recollection of intercourse, clarification of abnormal menses, inconsistent contraception use, and provider or patient request. The rate of hCG testing decreased each year, from 42.9% (15 of 35) in 2015 to 29.0% (95 of 328) in 2016, 22.0% (108 of 491) in 2017, and 16.5% (26 of 158) in 2018 ( Table 2 ). Rates of laboratory pregnancy testing ordered before the procedure visit similarly decreased. Table 2 PREG Results PREG items Calendar year Total (N=1012) 2015 (N=35) 2016 (N=328) 2017 (N=491) 2018 (N=158) I’ve had a negative pregnancy test today 1 (2.9%) 15 (4.6%) 18 (3.7%) 1 (0.6%) 35 (3.5%) Category A items  I am pregnant 0 0 0 0 0  I’ve had a bilateral tubal ligation 3 (8.6%) 5 (1.5%) 10 (2.0%) 0 (0.0%) 18 (1.8%)  I’ve had a hysterectomy or BSO 0 0 0 0 0  I am menopausal and >45 years old 1 (2.9%) 0 (0.0%) 1 (0.2%) 0 (0.0%) 2 (0.2%)  I have a current IUD in place 5 (14.3%) 56 (17.1%) 99 (20.2%) 40 (25.3%) 200 (19.8%)  I have a current contraceptive implant in place 0 (0.0%) 32 (9.8%) 27 (5.5%) 17 (10.8%) 76 (7.5%) Category B items  I have not had sexual intercourse with a man since the start of my last normal menstrual period 10 (28.6%) 103 (31.4%) 177 (36.0%) 59 (37.3%) 349 (34.5%)  My partner has had a vasectomy and he has had a negative post-surgery semen analysis 3 (8.6%) 18 (5.5%) 20 (4.1%) 7 (4.4%) 48 (4.7%)  I started bleeding from a normal period within the last 7 days 9 (25.7%) 105 (32.0%) 170 (34.6%) 53 (33.5%) 337 (33.3%)  I reliably use hormonal contraception 11 (31.4%) 118 (36.0%) 152 (31.0%) 44 (27.8%) 325 (32.1%) Category C items  I think I may be pregnant or would like a pregnancy test 0 (0.0%) 0 (0.0%) 0 (0.0%) 4 (2.5%) 4 (0.4%)  None of the above apply 6 (17.1%) 40 (12.2%) 46 (9.4%) 9 (5.7%) 101 (10.0%) Mutually exclusive categorization based on the hierarchy listed below (top to bottom)  I think I may be pregnant or would like a pregnancy test 0 (0.0%) 0 (0.0%) 0 (0.0%) 4 (2.5%) 4 (0.4%)  I had a negative pregnancy test today 1 (2.9%) 15 (4.6%) 18 (3.7%) 1 (0.6%) 35 (3.5%)  Marked an X in Category A only 5 (14.3%) 41 (12.5%) 73 (14.9%) 26 (16.5%) 145 (14.3%)  Marked an X in Category B only 21 (60.0%) 182 (55.5%) 292 (59.5%) 88 (55.7%) 583 (57.6%)  Marked an X in both Categories A & B 2 (5.7%) 50 (15.2%) 62 (12.6%) 30 (19.0%) 144 (14.2%)  None of the above apply 6 (17.1%) 40 (12.2%) 46 (9.4%) 9 (5.7%) 101 (10.0%) PREG Results Of women with a negative pregnancy test result, 158 (56.6% of all 279 pregnancy tests) had at least 1 group A or group B criterion to exclude pregnancy, including 51 who identified a current normal menses. Additionally, some women who identified their IUC or implant as expired were within the manufacturer’s guidelines for duration of use so actually had group A criteria in addition to their negative pregnancy test result. Other women with negative results had LARC that had been in place beyond the manufacturer’s guidelines so did not meet group A criteria but were within a duration of documented efficacy (eg, use of the levonorgestrel-releasing intrauterine system for less than 6 years or implant use just slightly longer than 36 months) for pregnancy exclusion. 12 , 13

The study was approved by the Mayo Clinic Institutional Review Board. The staff from the PCG clinic involved in the study included desk attendants who administer PREG, nurses who confirm PREG completion, and advanced practice providers (APPs; physician assistants, nurse practitioners) who review and discuss the PREG with the patient. The APPs from the PCG clinic have a special interest in women’s health and have additional training in gynecologic consultations and procedures. At the time of appointment scheduling, women were encouraged to continue efforts to prevent pregnancy until their appointment. Nursing counseling was available on request. An interprofessional team approach was used in the implementation of PREG. PREG ( Figure 1 ) was provided at the time of PCG clinic appointment check-in to women aged 18 to 50 years seen for contraceptive procedures (IUC and implant placement) from September 30, 2015, through April 30, 2018. Exclusion criteria were the inability to independently read English and complete PREG and denial of access to their medical records for research purposes. Figure 1 Pregnancy Reasonably Excluded Guide. Pregnancy testing was not required when responses were marked in category A or B. Pregnancy testing was required when responses were marked in category C. Used with permission of Mayo Foundation for Medical Education and Research. Pregnancy Reasonably Excluded Guide. Pregnancy testing was not required when responses were marked in category A or B. Pregnancy testing was required when responses were marked in category C. Used with permission of Mayo Foundation for Medical Education and Research. We also developed a demographic characteristics questionnaire ( Figure 2 ) to support clinic quality and flow, which was administered along with PREG. This questionnaire gathered patient characteristics including desired contraceptive procedure, menstrual history, prior contraception, history of gynecologic infection, gravidity and parity, fertility history, perimenopausal symptoms, and cervical cytological history. Figure 2 Demographic characteristics questionnaire. Used with permission of Mayo Foundation for Medical Education and Research. Demographic characteristics questionnaire. Used with permission of Mayo Foundation for Medical Education and Research. Nursing staff confirmed PREG completion and relayed the information to the APP. The PREG and demographic characteristics questionnaire were then reviewed and discussed in private with the patient. If hCG testing was indicated per APP order, nursing staff performed point-of-care urine pregnancy testing with the ICON 20 hCG (HemoCueAmerica), which has a threshold of 20 mIU/L in urine. Institutional quality control processes for this testing were continued. The contraceptive procedure was performed if appropriate. If the patient’s pregnancy status was uncertain, the visit was rescheduled. PREG contains 3 groups of statements that help determine whether pregnancy testing is required. Group A statements determine whether women are either pregnant or reasonably not pregnant that day and unlikely to become pregnant in the near future. Women who met criteria for group A did not require pregnancy testing. Group B statements determine whether women are reasonably not pregnant that day but could become pregnant in the near future. Women who met these conditions did not require pregnancy testing that day but could require pregnancy testing in the future. Group C statements identify women who state they might be pregnant, want a pregnancy test, or do not meet criteria in group A or B. Women in this category required preprocedure pregnancy testing. Women with a negative pregnancy test result and criteria from group A (except previously diagnosed pregnancy), group B, or both are reasonably not pregnant and can proceed with the scheduled procedure. Pregnancy testing could proceed or be added if either the patient or physician desires it during or after review of PREG criteria. Review and discussion of PREG criteria are absolute requirements for this process. For women in group C with a negative hCG result and no criteria in group A or B, pregnancy may be excluded or conception may have occurred in the preceding 12 to 13 days, depending on menstrual and coital history. 9 For example, consistent condom use or an interval of 17 days since unprotected coitus, along with a negative urine hCG result, were sufficient to reasonably exclude pregnancy and can be especially helpful criteria in women with irregular or long cycles. For women seeking removal and replacement of LARC after the US Food and Drug Administration–approved interval of use, the time since insertion was also considered. This determination was made by the APP during review and discussion of PREG and hCG results with the patient. PREG and demographic characteristics questionnaire results were stored in a database. Reviews of the electronic health record were used to gather information on age, race, date of procedure, whether hCG testing was completed along with the result, and whether the visit was rescheduled. These data were summarized for the study interval and subdivided by calendar year. These data were then compared with data collected for 185 women who underwent contraceptive procedures in the primary care setting in 2015 before PREG implementation. SQUIRE 2.0 guidelines were used as a framework for reporting findings. 10

Implementation of PREG in primary care for scheduled IUC or implant insertion at Mayo Clinic was associated with eligibility for scheduled same-day insertion in 96.4% overall and in 75.9% without the need for hCG testing. Use of PREG creates a standardized approach to screen for pregnancy status. It facilitates the delivery of comprehensive care while minimizing unnecessary hCG testing, the risk of false-positive results, delayed care, and cost. Negative hCG results require interpretation in the context of menstrual and coital history. PREG use is sustainable over time and generalizable across the Mayo Clinic health care system and beyond. Continued implementation of PREG in other settings will aid in time- and cost-efficient health care for reproductive-aged women.

Before the implementation of PREG, no formal process existed for pregnancy assessment, and hCG testing was based on physician discretion and practice style. PREG use in this primary care contraceptive procedure clinic setting allowed more women to undergo same-day contraceptive procedures. Reduced delay in initiating contraception/LARC placement is associated with lower unplanned pregnancy rates. 14 With PREG, pregnancy status was more consistently and objectively documented in this screening population, and no failure of pregnancy detection on procedure day was identified. PREG focuses pregnancy testing on women who have a higher likelihood of pregnancy. By identifying women with a very low likelihood of pregnancy using historical criteria, more women were able to forgo unnecessary pregnancy testing, thus decreasing delay in access to contraceptive procedures, rates of false-positive pregnancy test results, and personal and health care costs. PREG presents a structured opportunity for shared decision making regarding preprocedural pregnancy testing and had no unintended negative consequences related to use or cost in our setting. Rates of pregnancy testing, in fact, decreased. False-positive hCG results are more common in perimenopausal and menopausal women because of increased secretion of pituitary hCG. 15 Many perimenopausal woman use LARC because levonorgestrel IUC is efficacious in managing abnormal uterine bleeding; thus, pregnancy assessment and avoidance of unnecessary testing is important to avoid contraceptive or therapeutic delay. 16 False-positive urine hCG results have also been reported with adenomyosis and interfering rheumatoid factor. 17 , 18 Implementation of PREG for screening in the primary care setting creates a standardized approach to assess for pregnancy. PREG is also appropriate for use in noncontraceptive preprocedural settings. With continued use, physicians became more comfortable with its usefulness and validity. The rate of hCG testing performed decreased each year, from 42.9% in 2015 to 16.5% in 2018, which indicated an increased physician reliance on PREG. The testing rate of 42.9% in 2015 was higher than expected, but the downward trend through 2018 illustrated that practice change takes time. The latest hCG testing rate of 16.5% is similar to that seen with PREG use in the department of obstetrics and gynecology, 5 which supports that the decreased hCG testing was associated with PREG implementation. PREG has now been implemented successfully in several settings at Mayo Clinic. 5 , 6 Patient-physician review and discussion of pregnancy exclusion criteria remained important to improve the validity of the assessment and further ensure that the woman was a candidate for screening alone (ie, asymptomatic) vs requiring a diagnostic evaluation (eg, abnormal menses or acute pelvic pain). Additionally, the criteria review strengthened the context of negative hCG results and facilitated interpretation. Physician involvement in screening also helped improve accuracy, decrease overreliance on hCG testing, and improve patient understanding and acceptance. PREG enhanced quality patient care components, including safe and efficient patient-centered care with testing when beneficial, avoidance of testing when potentially detrimental, and contextual interpretation of negative hCG results. These characteristics associated with PREG benefit the patient and the health care system. This study reports on the generalizability and feasibility of PREG implementation in a primary care contraceptive procedure clinic setting. No ethical conflicts were apparent. PREG was administered to women who could read English and independently complete PREG. We did not perform PREG screening in women who could not read or independently complete the assessment to decrease the risk of misunderstanding or translation error. Further studies would need to be performed on the use of PREG in women who could not read or independently complete it. Results mirrored successful implementation in the gynecology specialty clinic at Mayo Clinic, and a higher PREG acceptability in the primary care setting may be due to previous physician experience with PREG. 5 The multiyear study was done to control for this effect of previous experience and to ensure longevity of PREG use. The rate of positive pregnancy tests (0.1%) was lower than the 2.6% to 7.95% positive rates reported in some family planning settings. 2 , 19 The benefit of a screening tool such as PREG may be lower in populations with higher pregnancy rates. The sample size of 1012 visits over several years is a study strength and documented the success of PREG implementation. The interprofessional team approach was especially effective because it provided a time-efficient process and consistent messaging to patients while maintaining overall patient flow. Women were not specifically evaluated for pregnancy after their procedure when a pregnancy present at the time of the procedure could be detected. However, women seen in this primary care clinic would have been unlikely to seek care for a pregnancy outside of the institution because of insurance coverage and care options. Additionally, a quality review would have been initiated for LARC placement during pregnancy. Before the opening of the PCG clinic and use of PREG, contraceptive procedures were done across the primary care practice by multiple physicians, with no consistent documentation. For this reason, the comparison data did not accurately capture the same-day access or rescheduling rates. Thus, estimated historical same-day access and rescheduling rates had to be determined by analysis of same-day eligibility rates for contraceptive procedure based on the criteria of current menses and IUC or implant in place responses. This is a study limitation. Women in these procedure appointments had typically identified the desire for IUC or implant insertion in advance. They may have received information from their care team or community contacts on circumstances that exclude pregnancy (appointment during menses) and avoidance of pregnancy before the scheduled appointment (contraceptive use or abstinence). Other women had LARC that was incorrectly identified by them as being past the US Food and Drug Administration–approved duration of use. Other LARC was due to be replaced but was within a time frame during which continued efficacy had been documented. 12 , 13 These factors most likely affected the high rate of procedures completed on the scheduled day.