Objective
The objective of the study was to evaluate perceived weight gain in women using contraception and determine the validity of self-reported weight gain.
Study Design
We analyzed data from new contraceptive method users who self-reported a weight change at 3, 6, and 12 months after enrollment. We examined a subgroup of participants with objective weight measurements at baseline and 12 months to test the validity of self-reported weight gain.
Results
Thirty-four percent of participants (1407 of 4133) perceived weight gain. Compared with copper intrauterine device users, implant users (relative risk, 1.29; 95% confidence interval, 1.10–1.51) and depot medroxyprogesterone acetate users (relative risk, 1.37; 95% confidence interval, 1.14–1.64) were more likely to report perceived weight gain. Women who perceived weight gain experienced a mean weight gain of 10.3 pounds. The sensitivity and specificity of perceived weight gain were 74.6% and 84.4%, respectively.
Conclusion
In most women, perceived weight gain represents true weight gain. Implant and depot medroxyprogesterone acetate users are more likely to perceive weight gain among contraception users.
Reported weight gain is one of the main reasons that women are not satisfied with their contraceptive method. In clinical studies evaluating the subdermal implant, 13.7% of users reported weight gain as an adverse effect of the method, and 2.3% reported weight gain as the reason for having their implant removed (Implanon package insert, 2006; Organon USA Inc, Roseland, NJ). However, perceived weight gain is rarely validated with objective measures.
The literature regarding contraception and objective measurements of weight gain has focused on certain methods, mainly oral contraceptives and depot medroxyprogesterone acetate (DMPA). Currently DMPA is the only contraceptive method consistently associated with significant weight gain. However, the literature fails to address how women perceive various contraceptive methods to affect their weight and whether their perception of weight gain is accurate.
Regardless of accuracy, a woman’s perception of her weight while using a contraceptive method may be just as important as the number on a scale, particularly regarding method satisfaction and continuation. In a study on self-weighing frequency and weight loss, 20% of adults at baseline reported they never weigh themselves, and 40% reported weighing themselves less than once a week. Therefore, many Americans rely on perceived weight as a method of weight assessment.
The purpose of this analysis was to evaluate perceived weight gain during the first 12 months after women started a new contraceptive method and to determine the validity of self-reported weight gain. To test the accuracy of perceived weight gain, we examined a subset of women who also received height and weight measurements at baseline and 12 months after enrollment.
Materials and Methods
This was a secondary analysis of the Contraceptive CHOICE Project. CHOICE is a prospective cohort study developed to promote the use of long-acting reversible contraception (LARC). The study provides contraception for 2-3 years at no cost to 9256 women in the St. Louis area in an effort to reduce unintended pregnancies in our region.
English- or Spanish-speaking women between 14 and 45 years of age were eligible for the study if they met the following criteria: (1) resided in the study’s designated recruitment region; (2) were willing to start a new method of reversible contraception or were currently not using any method; (3) did not desire to become pregnant for at least 12 months; and (4) were sexually active with a male partner or planning to become sexually active in the next 6 months. Women were excluded if they had a history of tubal sterilization or hysterectomy. Study enrollment began in August 2007 and ended in September 2011.
All enrolled participants received contraceptive counseling at one of the designated recruitment sites, which included the university-based clinic, 2 abortion clinics, and several community-based clinics. Counseling sessions at the university-based clinic presented women with all reversible contraceptive methods and their associated effectiveness, side effects, benefits, and risks to help participants make an informed decision. Counseling sessions at the remaining clinics were based on what was usual contraceptive counseling for that clinic location. In addition, during the baseline visit, a comprehensive assessment was performed to collect demographic and reproductive history information, a screening was conducted for sexually transmitted infections, and measurement of height and weight was done. The woman’s baseline body mass index (BMI) was calculated as weight (kilograms)/height (meters) 2 and categorized into 1 of the following BMI groups: underweight (<18.5 kg/m 2 ), normal (18.5-24.9 kg/m 2 ), overweight (25-29.9 kg/m 2 ), or obese (≥30 kg/m 2 ).
We performed telephone interviews using standardized survey instruments at 3 and 6 months after enrollment and every 6 months for the duration of the follow-up. Participants were compensated with a $10.00 gift card at each interview. During the follow-up surveys, participants were asked to self-report any changes in their weight. Specifically, the survey question asked, “Since we last spoke, has your weight changed by 5 pounds or more?” If the participant answered yes, she was further asked, “Did you (a) gain weight, (b) lose weight, (c) both gain and lose weight, or (d) you don’t know?” This question was not included until a revised version of the survey was introduced on July 1, 2008. Therefore, only participants who answered this question are included in this analysis.
For this analysis, we evaluated women’s perceived weight change during the first 12 months of starting a new contraceptive method using data from their 3, 6, and 12 month surveys. A participant was excluded from the analysis if we could not reach a definite conclusion regarding her direction of perceived weight change over 12 months. For example, a participant was excluded if she was missing a response to the question above in any of the 3, 6, or 12 month surveys or had responded yes to weight change but was missing a response to how her weight had changed (ie, weight gain or loss). In addition, if a participant ever reported both weight gain and weight loss or she did not know the answer to the question, “How did your weight change?” in any of the 3, 6, or 12 month follow-up surveys, she was excluded from the analysis sample because we could not determine the direction (gain or loss) of her weight change. Finally, if a participant’s combined responses of all 3 surveys included any combination of gain and loss, she was excluded because we could not determine her net sum of perceived weight change because we had no way of knowing which weight change was the most prominent.
We defined 3 groups: (1) perceived weight gain; (2) perceived no weight change; and (3) perceived weight loss. Group 1 (perceived weight gain) included any participant who reported weight gain in at least one of her 3, 6, or 12 month surveys and reported no change at all of the remaining 3, 6, or 12 month survey(s). Group 2 (perceived no weight change) included participants who, at all of the 3, 6, and 12 month follow-up surveys, responded, “No, my weight has not changed by 5 pounds or more since we last spoke.” Group 3 (perceived weight loss) included any participant who reported weight loss in at least 1 of her 3, 6, or 12 month surveys and reported no change at all of the remaining 3, 6, or 12 month survey(s).
All women enrolled in CHOICE on or after June 15, 2010, were offered eligibility screening and, if eligible, enrolled in a separate substudy to objectively assess 12 month weight change with progestin-only methods or the copper intrauterine device (IUD). Eligibility for this substudy included the following: (1) CHOICE enrollment at the university-based clinic to retain consistency in the scale used to assess weight; (2) age of 18 years or older; (3) using the levonorgestrel intrauterine system (LNG-IUS), copper T380A IUD, subdermal implant, or DMPA as the current method; (4) continued use of 1 of the previously mentioned methods for at least 11 months; and (5) a willingness to return to the university-based clinic for a 12 month visit for an objective weight assessment.
At the time of data analysis, 415 women had enrolled and completed this substudy of objectively measured weight change. We calculated the participant’s observed weight change by subtracting her recorded weight at baseline from the recorded weight at her 12 month follow-up. The observed weight change was classified as 1 of the following: a gain if the calculated weight change was a 5 pound or greater increase; no change if the calculated weight change was less than a 5 pound change in either direction; and a loss if the calculated weight change was a 5 pound or greater decrease. We chose a 5 pound change to be consistent with our survey instrument.
The observed weight change categories of gain, no change, and loss were then compared with the perceived weight change groups (1, 2, and 3) as previously defined. To estimate sensitivity, specificity, agreement, and relative risk calculations, we combined group 2 (perceived no weight change) and group 3 (perceived weight loss) to create a dichotomous variable for perceived weight gain: yes (perceived weight gain) or no (no perceived weight gain). We created a similar variable for objective weight gain, in which the no-weight-gain group included the measured no-change and weight-loss groups.
The CHOICE protocol and substudy described in the previous text were approved by the Washington University in St. Louis School of Medicine Human Research Protection Office. The methodic details of the Contraceptive CHOICE Project have been published in a separate report.
Statistical Analyses
We compared the baseline demographic and behavioral characteristics among the 3 perceived weight change groups (gain, no change, and loss) in both the CHOICE analysis sample and our substudy sample using χ 2 or Fisher exact test where appropriate. We compared the mean objective weight change of the 3 perceived weight gain groups in the substudy using a 1-way ANOVA test. We calculated a kappa statistic to evaluate the agreement between objective and perceived weight gain. We calculated the sensitivity and specificity of perceived weight gain compared with the gold standard of objectively measured weight gain.
We also calculated the relative risk (RR) of perceived weight gain by contraceptive methods and 95% confidence intervals (CIs) from Poisson regression using the SAS GENMOD procedure (SAS Institute, Cary, NC). This approach provides an unbiased estimate of the relative risk in the case of common outcomes (>10%). Confounding was defined as a greater than 10% relative change in the association between perceived weight gain and method choice with or without the potential confounding factor in the model. Confounders were included in the final model to obtain an adjusted RR of perceived weight gain among contraceptive methods. All statistical analysis was performed using SAS 9.2 software. The significance level alpha was set at 0.05.
Statistical Analyses
We compared the baseline demographic and behavioral characteristics among the 3 perceived weight change groups (gain, no change, and loss) in both the CHOICE analysis sample and our substudy sample using χ 2 or Fisher exact test where appropriate. We compared the mean objective weight change of the 3 perceived weight gain groups in the substudy using a 1-way ANOVA test. We calculated a kappa statistic to evaluate the agreement between objective and perceived weight gain. We calculated the sensitivity and specificity of perceived weight gain compared with the gold standard of objectively measured weight gain.
We also calculated the relative risk (RR) of perceived weight gain by contraceptive methods and 95% confidence intervals (CIs) from Poisson regression using the SAS GENMOD procedure (SAS Institute, Cary, NC). This approach provides an unbiased estimate of the relative risk in the case of common outcomes (>10%). Confounding was defined as a greater than 10% relative change in the association between perceived weight gain and method choice with or without the potential confounding factor in the model. Confounders were included in the final model to obtain an adjusted RR of perceived weight gain among contraceptive methods. All statistical analysis was performed using SAS 9.2 software. The significance level alpha was set at 0.05.
Results
Of the 7977 CHOICE participants who reached 12 months at the time of data analysis, 4133 women met our inclusion criteria. Among the 3844 women excluded, 2968 had missing data regarding the weight question or the specific direction of weight change and 1146 reported both weight gain and loss. We found no significant differences in demographic and behavioral characteristics between the analysis sample and the women who were excluded because of inclusion criteria or inability to reach a definite conclusion regarding their weight change. Of the 4133 women in the analysis sample, 1407 (34.0%) perceived weight gain, 1634 (39.5%) perceived no weight change, and 1092 (26.4%) perceived weight loss.
Table 1 displays the demographic and behavioral characteristics of the analysis sample and the 3 perceived weight change groups: gain, no change, and loss. Women who perceived weight gain were significantly more likely to be African American, parous, uninsured, and less educated. The participants in the weight-gain group were also more likely to have trouble paying for basic necessities and receive public assistance. General health was highest among those who reported no perceived weight change. Compared with women who were normal or underweight at baseline (n = 1882), women who were overweight or obese at baseline (n = 2251) were more likely to report perceived weight gain (36.7% vs 30.9%) and weight loss (30.5% vs 21.5%) and less likely to report no weight change (32.8% vs 47.6%).
| Characteristics | Analysis sample (n = 4133) | Perceived weight gain (n = 1407) | Perceived no weight change (n = 1634) | Perceived weight loss (n = 1092) | P value a |
|---|---|---|---|---|---|
| Age, y b | .042 | ||||
| <21 | 851 (20.6) | 300 (21.3) | 342 (20.9) | 209 (19.1) | |
| 21-25 | 1624 (39.3) | 511 (36.3) | 670 (41.0) | 443 (40.6) | |
| >25 | 1658 (40.1) | 596 (42.4) | 622 (38.1) | 440 (40.3) | |
| Race b | < .001 | ||||
| Black | 1989 (48.1) | 823 (58.5) | 725 (44.4) | 441 (40.4) | |
| White | 1810 (43.8) | 472 (33.5) | 778 (47.6) | 560 (51.3) | |
| Other | 334 (8.1) | 112 (8.0) | 131 (8.0) | 91 (8.3) | |
| Marital status b | .011 | ||||
| Single | 2432 (58.9) | 845 (60.1) | 972 (59.6) | 615 (56.3) | |
| Living with partner | 900 (21.8) | 289 (20.6) | 345 (21.1) | 266 (24.4) | |
| Married | 551 (13.3) | 169 (12.0) | 234 (14.3) | 148 (13.5) | |
| Separated/divorced/widowed | 247 (6.0) | 103 (7.3) | 81 (5.0) | 63 (5.8) | |
| Insurance b | < .001 | ||||
| None | 1668 (40.6) | 641 (45.8) | 621 (38.2) | 406 (37.3) | |
| Private | 1873 (45.5) | 535 (38.3) | 806 (49.6) | 532 (48.9) | |
| Public | 571 (13.9) | 222 (15.9) | 199 (12.2) | 150 (13.8) | |
| Education b | < .001 | ||||
| High school diploma/GED or less | 1363 (33.0) | 546 (38.8) | 490 (30.0) | 327 (30.0) | |
| Some college | 1662 (40.2) | 572 (40.7) | 639 (39.1) | 451 (41.3) | |
| College degree or graduate school | 1107 (26.8) | 288 (20.5) | 505 (30.9) | 314 (28.7) | |
| Trouble paying for basic expenses b | < .001 | ||||
| No | 2608 (63.2) | 771 (54.9) | 1138 (69.8) | 699 (64.0) | |
| Yes | 1519 (36.8) | 633 (45.1) | 493 (30.2) | 393 (36.0) | |
| Receiving public assistance b | < .001 | ||||
| No | 2680 (64.9) | 816 (58.1) | 1136 (69.6) | 728 (66.7) | |
| Yes | 1448 (35.1) | 589 (41.9) | 496 (30.4) | 363 (33.3) | |
| Parity b | < .001 | ||||
| Nulliparous | 2030 (49.1) | 608 (43.2) | 867 (53.1) | 555 (50.8) | |
| Parous | 2103 (50.9) | 799 (56.8) | 767 (46.9) | 537 (49.2) | |
| Baseline BMI, kg/m 2 b | < .001 | ||||
| Underweight (<18.5) | 211 (5.1) | 45 (3.2) | 125 (7.7) | 41 (3.8) | |
| Normal (18.5-24.9) | 1671 (40.4) | 537 (38.2) | 770 (47.1) | 364 (33.3) | |
| Overweight (25-29.9) | 1040 (25.2) | 376 (26.7) | 360 (22.0) | 304 (27.8) | |
| Obese (≥30) | 1211 (29.3) | 449 (31.9) | 379 (23.2) | 383 (35.1) | |
| Baseline method b | < .001 | ||||
| LNG-IUS | 1892 (45.8) | 650 (46.2) | 705 (43.2) | 537 (49.2) | |
| Copper IUD | 532 (12.9) | 156 (11.1) | 212 (13.0) | 164 (15.0) | |
| Implant | 651 (15.8) | 266 (18.9) | 233 (14.3) | 152 (13.9) | |
| DMPA | 282 (6.8) | 130 (9.3) | 102 (6.2) | 50 (4.6) | |
| Pills, patch, ring | 774 (18.7) | 204 (14.5) | 381 (23.3) | 189 (17.3) | |
| General health (baseline) b | < .001 | ||||
| Excellent | 1199 (29.1) | 352 (25.1) | 545 (33.4) | 302 (27.7) | |
| Very good | 1681 (40.7) | 549 (39.2) | 659 (40.4) | 473 (43.3) | |
| Good | 1016 (24.6) | 399 (28.4) | 365 (22.3) | 252 (23.1) | |
| Fair | 215 (5.2) | 94 (6.7) | 61 (3.7) | 60 (5.5) | |
| Poor | 15 (0.4) | 8 (0.6) | 3 (0.2) | 4 (0.4) | |
| General health (12 months) b | < .001 | ||||
| Excellent | 1012 (24.6) | 267 (19.0) | 475 (29.2) | 270 (24.8) | |
| Very good | 1651 (40.1) | 522 (37.2) | 689 (42.4) | 440 (40.3) | |
| Good | 1115 (27.1) | 433 (30.9) | 383 (23.6) | 299 (27.4) | |
| Fair | 322 (7.8) | 167 (11.9) | 75 (4.6) | 80 (7.3) | |
| Poor | 19 (0.4) | 14 (1.0) | 3 (0.2) | 2 (0.2) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
