Abstract
Background:
Women with gynecologic disorders requiring a hysterectomy often have co-existing psychiatric diagnoses. A change in the dispensing pattern of antidepressant (AD) and antianxiety (AA) medications around the time of hysterectomy may be due to improvement in gynecologic symptoms, such as pelvic pain and abnormal bleeding, or the emotional impact of the hysterectomy. Unfortunately, these dispensing patterns before and after hysterectomy are currently undescribed.
Objectives:
To model the dispensing patterns of AD and AA medications over time among women with psychiatric disorders before and after benign hysterectomy for endometriosis and uterine fibroids; and to characterize clusters of patients with various dispensing behaviors based on these patterns.
Design:
Retrospective cohort study.
Methods:
This is a study of women who underwent a benign hysterectomy using data from the Merative MarkertScan® Research Databases (Ann Arbor, MI, USA). Inclusion criteria were reproductive-aged women (18–50 years), diagnosis of at least one mood or anxiety disorder, and at least one dispensing of AD or AA medications. We measured monthly adherence and persistence of AD/AA medication use over 12 months after hysterectomy. Group-based-trajectory modeling (GBTM) was used to identify trajectory groups of monthly AD/AA medication dispensing over the study period. Multinomial logistic regression was used to identify factors independently associated with individual dispensing trajectory patterns.
Results:
For a total of 11,607 patients, 6 dispensing trajectory groups were identified during the study period: continuously high (27.0%), continuously moderate (21.9%), continuously low (17.9%), low-to-high (10.0%), moderate-to-low (9.8%), and low-to-moderate (13.4%). Compared with the continuously high group, younger age, no history of a mood disorder, and uterine fibroids were clinical predictors of low dispensing. The discontinuation rate at 3 months after hysterectomy was higher at 88.6% in the continuously low group and at 66.5% in the continuously low-to-moderate group.
Conclusions:
This study demonstrates that GBTM identified six distinct trajectories of AD/AA medication dispensing in the perioperative period. Trajectory models could be used to identify specific dispensing patterns for targeting interventions.
Plain language summary
Women with gynecologic disorders often have coexisting psychiatric diagnoses. A change in the dispensing pattern of antidepressant and antianxiety medications may be due to improvement in gynecologic symptoms or the emotional impact of the hysterectomy. However, static measures, such as the proportion of days covered or medication possession ratio, may not adequately predict meaningful dispensing patterns. Using the group-based trajectory modeling, 6 distinct patterns of medication dispensing over the perioperative periods of women with benign hysterectomy are identified and therefore used to assess how certain clinical characteristics influence these dispensing patterns. This study concludes that trajectory modeling may be a more appropriate approach to investigating dispensing patterns among women with preexisting psychiatric conditions.
Introduction
Hysterectomy is the most common major gynecologic surgery in the United States (US), with more than 60,000 procedures performed annually, 1 and often for benign indications, such as chronic pelvic pain and abnormal uterine bleeding secondary to endometriosis and uterine fibroids, that can have serious psychological sequalae in women. 2 Endometriosis, for example, has been reported to be associated with high rates of psychiatric morbidities, namely depression, anxiety, and overall poor quality of life. 3 Although it is conceivable that treatment of underlying gynecologic pathologies can lead to improvement in psychiatric symptoms, the evidence remains controversial as to whether that notion is true.4,5
From a different perspective, the direct association between hysterectomy and mental health has been historically described in the literature, predating to as early as the 1960s. In his study in 1968, Barker concluded that 7% of women undergoing a hysterectomy were referred to a psychiatrist within a mean duration of 4.5 years postoperatively, a rate that is 2.5 and 3 times higher than in women undergoing other procedures and in the general population, respectively. 6 This notion is further supported by a contemporary Australian study showing that women having a hysterectomy are at a significantly higher risk of new depressive symptoms postoperatively. 7 On the other hand, a meta-analysis suggests that benign hysterectomy may not be adversely associated with but rather improve postoperative psychiatric morbidity, 5 perpetuating controversies around the true impact of hysterectomy on women’s psyche.
Remission and relapse of psychiatric disease, albeit extremely important measures of disease burden, remain difficult to assess objectively. 8 Assessing psychiatric medication use has proved beneficial as a surrogate for mental illness burden9,10 and could therefore provide quantifiable information on psychiatric disease progression or regression. While several studies have explored the association between hysterectomy and psychiatric disease, changes in the patterns of psychiatric mediation use post-hysterectomy have not been studied. In administrative claims data, medication dispensing is often assessed through measures such as the proportion of days covered (PDC).11,12 However, patients with the same PDC-based adherence measure can exhibit very different patterns of adherence and the differences may be blurred when they are lumped together. To help define a dynamic dispensing pattern of medications, we applied group-based trajectory modeling (GBTM) for summarizing long-term medication adherence and accounting for the nature of adherence over time. The advantage of this method over PDC is to identify individuals who had similar longitudinal dispensing patterns of medications and to reveal the change in medication refill behavior over time for each group. This method has been previously used to study medication dispensing patterns.13,14 Using an administrative claims database and GBTM, we aim in this study to examine the dynamic patterns of antidepressant (AD) and antianxiety (AA) medication dispensing after benign hysterectomy for endometriosis and uterine fibroids among women with preexisting psychiatric disease and identify the independent predictors of these patterns.
Materials and methods
Data source
This study utilized data from the Merative MarketScanⓇ Commercial Claims (Ann Arbor, MI, USA) and Encounters database, a validated national commercial insurance repository gathering prescription information for patients enrolled in private health plans. 15 It includes >263 million covered individuals, 40 health plans, 350 unique carriers, and medical claims for inpatient and outpatient services and pharmacy claims across the United States. The study was exempt by the Johns Hopkins University Institutional Review Board as the database contains only de-identified data.
Study design and patient cohort
In this retrospective cohort study, we identified reproductive-aged women (18–50 years) who underwent a benign hysterectomy for endometriosis or uterine fibroids from January 2011 to December 2016 (>90% of all benign hysterectomies for reproductive-aged women) and had at least one dispensing of AD or AA medications in the preoperative period. We applied the Current Procedural Terminology (CPT) codes to find hysterectomies (Appendix). The index date was defined as the date of hysterectomy. To verify AD and AA medication use in preoperative and postoperative periods, patients had to have a diagnosis of at least one mood or anxiety disorder (Appendix), prescription coverage, and be continuously enrolled in the database for at least 1 year both prior to and after hysterectomy (Figure 1(a)). We excluded patients diagnosed with schizophrenia, bipolar disorder, and psychotic disorders (paranoid states) (Appendix). AA or AD medication dispensing was evaluated by PDC. Patients were excluded if their pharmacy claim record or index diagnosis record was missing. Lastly, we excluded patients with radical hysterectomy, gynecologic malignancy within 1 year before or after hysterectomy, a delivery within 12 weeks prior to hysterectomy, and those with >1 procedure type on the index date (Appendix). Power analysis for calculating sample size was not conducted in this study. 16 The Strengthening Reporting of Observational studies in Epidemiology guidelines were followed for constructing the manuscript. 17
(a) Flowchart of the study population selection process. The number of subjects is included in the flowchart. (b) Schematic presentation of the study timeline and AD/AA medication use.
AD and AA medication dispensing
AD and AA medication dispensing was determined utilizing prescription drug codes in the MarketScanⓇ database. To identify AD and AA medications, therapeutic classes (THERCLS) and subclasses of generic drug identifiers were grouped into seven categories: selective serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors (SNRIs), benzodiazepines, ADs, anticonvulsant mood stabilizers, tricyclic antidepressants (TCAs/TeCAs), and other (lithum, monoamine oxidase inhibitors) (Appendix). For women with multiple AD and AA medication prescriptions, only the longer days’ supply prescription was included in the analysis.
Outcome
Our primary outcome was the trajectory of dispensed AD and AA medications for women with a mood or anxiety disorder who underwent a benign hysterectomy. To assess monthly PDC trajectories, we first calculated monthly PDC for AD or AA medications for each patient. As long-term medications were prescribed from 3 to 6 months’ supply, monthly PDC tends to be lower from 12 months to 6 months prior to hysterectomy. Thus, we evaluated the dispensing period from 6 months before to 12 months after hysterectomy. Monthly PDC is calculated by dividing the total number of days covered by AD or AA medications in each 30-day period by 30. Thus, the PDCs can range from 0 (0 days covered/30 days) to 1 (30 days covered/30 days). PDCs were calculated using prescription fill date and number of days’ supply. 12 Each month, PDC was dichotomized using a cut-off value of ⩾0.8 to define adherence in each month (Figure 1(b)).18,19 Next, we used GBTM to group patients into different trajectories of dispensed AA/AD medications (see the statistical analysis section for more details). 13 GBTM, a type of mixed model originally developed to model changes in behavior, 20 will help summarize the long-term medication adherence and identify individuals who had similar longitudinal medication dispensing patterns, revealing changes in medication refill behavior over time for each group. This method has been previously used to study medication dispensing patterns.13,14 In previous studies, GBTM has demonstrated improved characterization of medication refill in comparison with standard adherence-refill metrics such as PDC.21,22
Covariates (predictors of dispensed AD/AA medications trajectories)
Variables assessed as potential predictors of dispensing trajectories of AD and AA medications were demographic data, comorbidities, including preoperative pain diagnosis (back or neck pain, fibromyalgia) and anemia, psychiatric diagnosis (mood disorder, anxiety, insomnia), and gynecologic disorders, including inflammatory disease of female pelvic organs, prolapse of female genital organs, and menstrual disorders, and Charlson comorbidity index. 23 Preoperative and early postoperative factors included hysterectomy approach (abdominal, laparoscopic, or vaginal), surgical indication (endometriosis and uterine fibroids), length of inpatient hospital stay (days), and surgical complications within 30 days after the index date. Patient demographics, clinical characteristics, and the MarketScanⓇ codes by which they were obtained can be found in the Appendix.
Statistical analysis
Using SAS software (version 9.4; SAS Institute Inc., Cary, NC, USA), we applied GBTM to group patients according to their dispensing patterns of AD and AA medications. This method was implemented using the SAS “Proc traj” add-on package, which is a specialized mixture model that estimates multiple groups within the study population, (http://www.andrew.cmu.edu/user/bjones/index.htm). GBTM simultaneously estimates several regression models using a maximum likelihood approach that combines data from all models. 20 We used a multinomial logistic regression model with an intercept to estimate the probability of belonging to each potential trajectory group according to patients’ dispensing patterns over time.13,24 Within each dispensing group, AD and AA medication use (PDC ⩾ 0.8) in each month was modeled as a smooth function of time using up to a third-degree polynomial function. GBTM outputs estimated probabilities of group membership for each patient and the dispensed trajectory curve for each group. In GBTM, dependent variables were the 18 monthly binary indicators of AD and AA medication use. We selected the number of possible dispensing groups by taking into account the proportion of estimated groups (smallest group at least 5% of patients) and using the Bayesian Information Criterion (BIC) statistic.20,25,26 After obtaining the number of trajectories, we varied the order of the polynomial function with liner, cubic, and quartic orders to describe the shape of the trajectories. We found that a third-degree polynomial function provided the best fit for modeling the probability of AD and AA medication dispensing. Based on these models, we predicted the probability of belonging to each group for patients and assigned patients to the trajectory group where they have the highest predicted probability of membership. 13
A multinominal logistic regression analysis was conducted to identify independent predictors of each trajectory of dispensed AD and AA medications in the study period. All covariates mentioned above in the year before hysterectomy as well as the patient’s trajectory were considered to derive the adjusted odd ratios in the model. Failure time analysis using the Kaplan–Meier method with log-rank test was conducted to evaluate the time to discontinuation and reinitiation of AD or AA treatment after discontinuation. The discontinuation was defined as a gap to refill within 45 days of end of days’ supply of the previous dispensing in the postoperative period. 27 The duration of AD or AA treatment was defined as from the date of hysterectomy to the discontinuation date of the treatment or was censored if it continued until 365 days after hysterectomy. 27 The analyses were performed using SAS version 9.4, (SAS Institute, Cary, NC, USA) and statistical significance was defined as p < 0.05.
Results
Study cohort demographics
From 2011 to 2016, 32,513 women met the inclusion criteria with at least one mood or anxiety disorder and continuous enrollment for 1 year before and 1 year after hysterectomy. After excluding noneligible women, the final cohort consisted of 11,067 patients (Figure 1(a)). The sociodemographic, operative, and diagnostic characteristics are described (Table 1 and Supplemental material Table S1). The mean age of the study cohort was 42.0 (±5.37) years. Regarding surgical indication, 54.7% of patients had endometriosis while 45.3% had uterine fibroids.
Characteristics of the study population according to each trajectory group (N = 11,067).
SD: standard deviation.
Prescription patterns of therapeutic class medication
We identified prescription patterns of different AD and AA drug classes: SSRIs, SNRIs, benzodiazepines, anticonvulsant, antidepressants, TCAs/TeCAs, and Other (lithium, monoamine oxidase inhibitors) over the study years. SSRIs and benzodiazepines were the most prescribed AD and AA medications. Of note, prescriptions patterns were similar across THERCLS over the study duration (Supplemental material Figure S1).
Dispensing pattern trajectories
In the trajectory analysis, patients were categorized in different iterations into 2, 3, 4, 5, 6, or 7 groups with similar dispensing trajectories. The six-group trajectory model was found to fit the data most accurately with the smallest value of BIC and the minimum 5% of subgroup sizes. Details of the trajectory model with BIC are presented (Supplemental material Table S2 and Supplemental material Figure S2). In this model, the six-group trajectory represented six dispensing patterns: continuously high (CH, 27.0%: 4), continuously moderate (CM, 21.9%: 3), continuously low (CL, 17.9%: 5), low-to-high (LH, 10.0%: 2), moderate-to-low (ML, 9.8%: 6), and low-to-moderate (LM, 13.4%: 1) (Figure 2).
Trajectory model for six-group dispensing patterns from 6 months before and 12 months after hysterectomy. For the six-group model, group 1 = low to moderate; group 2 = low to high; group 3 = continuously moderate; group 4 = continuously high; group 5 = continuously low; and group 6 = moderate to low.
The PDC in the postoperative period was 0.4 in LM, 0.4 in LH, 0.8 in CM, 0.9 in CH, 0.2 in CL, and 0.6 in ML groups. There was a statistically significant difference between all six groups (p < 0.0001, Table 1), emphasizing the importance of categorizing dispensing patterns.
Predictors of dispensed AD/AA medication
The demographic and clinical characteristics of patients in the six trajectory groups are presented in Table 1. After conducting a multinominal regression analysis, we identified different clinical predictors associated with dispensing patterns (Table 2). The most predictive factors of low dispensing compared with CH dispensing were age, no history of mood disorder, and uterine fibroids. Other factors predictive of low dispensing were no history of fibromyalgia, no history of insomnia, and history of anemia (Table 2).
Predictors of AD/AA medication dispensing associated per trajectory group.
Odds ratios and 95% CIs are adjusted for baseline covariates. ORs >1 are predictive of nonadherence.
p < 0.05. **p < 0.0001.
Discontinuation and reinitiation of the treatment
The discontinuation rate of AD/AA medications in the postoperative period was different in-between groups (p < 0.0001) and higher among ML, CL, and LM compared with other groups (Figure 3(a)). Using a Kaplan–Meier curve analysis, the 6-month probability of discontinuation was 5.9% for CH, 20.5% for LH, 31.0% for CM, 78.5% for ML, 79.3% for LM, and 96.6% for CL (Figure 3(a) and Supplemental material Table S3).
(a) % of patients who discontinued treatment (more than the allowable gap 45 days) for six-group at 3, 6, 9, and 12 months after hysterectomy. For the six-group model, group CH = “Continuously high,” CM = “Continuously moderate,” LH = “Low to high,” ML = “Moderate to low,” LM = “Low to moderate,” CL = “Continuously low.” p < 0.0001 comparing the six-group based on GBTM (log-rank test). (b) Time to re-initiation of the treatment after discontinuation for six-groups at 3, 6, 9, and 12 months after Hysterectomy. For the six-group model, group CH = “Conti. high,” CM = “Conti. moderate,” LH = “Low to high,” ML = “Moderate to low,” LM = “Low to moderate,” CL = “Conti. low.” p < 0.0001 comparing the six-group based on GBTM (log-rank test).
The proportion of patients who reinitiated AD/AA medications after discontinuation was higher among the ML, CL, and LM groups compared with other groups. We found that more than half of patients (50.1% in CL and 58.3% in LM) reinitiated their medications at 6 months after hysterectomy (Figure 3(b)). The median time to reinitiate AD/AA medications from the index date was 172 days in the CH group, 167 days in the CM group, 151 days in the ML group, 188 days in LH group, 168 days in the LM group, and 141 days in the CL group (Supplemental material Table S3).
Discussion
This study characterized the dispensing of AD and AA medications around the time of benign hysterectomy in women with preexisting psychiatric disease and identified six distinct dispensing trajectories ranging from CH to continuously low dispensing. It also identified younger patient age, absence of mood disorder, fibromyalgia, and insomnia, and surgical indication of uterine fibroids as independent predictors of low dispensing patterns. The absence of anxiety and other factors are not associated with dispensing patterns in the adjusted multivariate model (not statistically significant). The identification of a full range of dispensing patterns after benign hysterectomy suggests that GBTM can be more informative and specific than a single PDC aggregate estimate. In addition, if perceived as surrogates for burden of mental illness, these patterns indicate that the impact of hysterectomy on mental health can be multidirectional and multifactorial.
These distinct patterns can have different interpretations. For example, previous studies have suggested that depressive symptoms can improve after hysterectomy, 28 which can be attributed to relief from gynecologic symptoms that led to a hysterectomy, such as abnormal uterine bleeding and pelvic pain, and potentially to lower AD medication dispensing. Further, evidence has shown that sexual health overall improves after a hysterectomy,29,30 resulting in better quality of life and psychological outcomes. 31 By contrast, a group of other studies found a higher risk of psychiatric morbidity post-hysterectomy. Wilson et al concluded that women undergoing a hysterectomy, with or without ovarian conservation, had a significantly higher risk of new depressive symptoms over a 12-year follow-up period. 7 Likewise, Korean women with uterine fibroids had an increased burden of mood disorders after hysterectomy compared with women having a uterus-conserving surgery, 32 which could be related to altered self-perception, feelings of decreased femininity, and lower self-esteem. 33
PDC is commonly used and a single aggregate estimate; however, by quantifying various dispensing patterns to a single ratio, PDC collapses a broad spectrum of adherence behaviors thus, masking various longitudinal dispensing patterns of the individuals’ behavior. 34 When compared with dichotomous PDC, GBTM shows that patient dispensing patterns are varied over time depending on an individual’s medical status, comorbidities, or many other clinical factors. Indeed, a patient with PDC of 0.4 has been grouped into two different dispensing patterns: (1) low to moderate and (2) low to high over the follow-up period. The outcome of trajectory groups with GBTM suggests heterogeneity among patients that requires a more sophisticated array of interventions to improve medication adherence. However, limitation lies in the GBTM could be influenced by this heterogeneity, such as the length of study evaluation (12–36 months, etc.,), the number and types of AD/AA medications evaluated, practice setting, patient demographics, and more. Any of these variables can bias the outcome in a way that masks meaningful results.
While the above-mentioned findings by prior studies may initially seem conflicting, possibly due to discrepancies in study designs, length of follow-up period, and cultural factors unique to each study population, our modeling indicates that the impact of hysterectomy on psychiatric morbidity is not necessarily unidirectional to begin with even in a single cohort of women. We were able to identify several independent factors associated with specific patterns of psychiatric medication dispensing after the procedure. Our finding of an association between uterine fibroids as a surgical indication and low dispensing patterns postoperatively was not in line with previous studies. 35 While definitive treatment of fibroids may alleviate a large burden of psychological symptoms, the profound impact of fibroids on mental health may linger despite their removal.35,36 Studies investigating medication compliance in patients with major depression reported that younger patients had higher drop-out rates and lower adherence to medications, 37 which comes in accordance with our finding of low dispensing patterns among younger women. Absence of a pain disorder or fibromyalgia was likewise associated with low dispensing in the LM and CL groups as AD and AA medications are often prescribed for such nonpsychiatric indications rather than a direct impact of hysterectomy.
Several clinical implications could be inferred from our study findings. We demonstrated that 66.5% of patients in the LM group have discontinued their AD and AA medications at 3 months after surgery, while 58.3% have reinitiated their treatment at 6 months after surgery. Monthly adherence (PDC) has increased from 30% at 3 months to 65% at 12 months after surgery in the LM group. Depression was found to be a strong independent predictor of postoperative morbidity, 38 and while depressive symptoms may appear to initially improve postoperatively, discontinuation of psychiatric medications should not be prematurely pursued and rather cautiously practiced, preferably in collaboration with mental health specialists. Women undergoing hysterectomy, especially those with preexisting psychiatric disorders, should undergo a thorough psychological evaluation preoperatively and be provided with the appropriate mental health resources, counseling, and postoperative expectations. Postoperatively, they should be closely monitored for relapse or worsening of their symptoms.
The results of this study should be interpreted with several limitations, especially those inherent to retrospective study designs. Our study population includes insured patients mostly from large employers, rendering our results possibly not generalizable to all women. Further, patient selection depends on the accuracy of diagnosis codes reported in claims data, and potential coding errors can lead to misidentification of disorders. However, we required patients to have at least one depressive or anxiety disorder in addition to at least one AD or AA prescription and claim to mitigate these biases. Dispensing measured as PDC ⩾0.8 as a surrogate for treatment adherence is not optimal for all patients, but using a cut-off value to classify patients into good or poor adherence groups is a common approach and served our purposes to assess dispensing patterns. 39 This method relies on prescription fulfillment rather than on patients taking the medication. We also could not identify whether medication discontinuation or reinitiation was patient or physician driven. On the other hand, our study was first to describe psychiatric medication dispensing patterns as a surrogate for burden of mental health after benign hysterectomy. In addition, our large population size and inclusive demographic characteristics, clinical data, and pharmacy claims data provide a more comprehensive picture of the dispensing patterns of AD and AA medications around the time of hysterectomy.
Conclusions
We identified six distinct trajectories of AD and AA medication dispensing after benign hysterectomy, representing the diverse impact of the procedure on mental health. Our trajectory models suggested that patients should be carefully monitored for psychiatric symptom progression postoperatively.
Supplemental Material
sj-docx-1-whe-10.1177_17455057241272218 – Supplemental material for Dispensing patterns of antidepressant and antianxiety medications for psychiatric disorders after benign hysterectomy in reproductive-age women: Results from group-based trajectory modeling
Supplemental material, sj-docx-1-whe-10.1177_17455057241272218 for Dispensing patterns of antidepressant and antianxiety medications for psychiatric disorders after benign hysterectomy in reproductive-age women: Results from group-based trajectory modeling by Ryota Ishiwata, Abdelrahman AlAshqar, Mariko Miyashita-Ishiwata and Mostafa A Borahay in Women’s Health
Supplemental Material
sj-docx-2-whe-10.1177_17455057241272218 – Supplemental material for Dispensing patterns of antidepressant and antianxiety medications for psychiatric disorders after benign hysterectomy in reproductive-age women: Results from group-based trajectory modeling
Supplemental material, sj-docx-2-whe-10.1177_17455057241272218 for Dispensing patterns of antidepressant and antianxiety medications for psychiatric disorders after benign hysterectomy in reproductive-age women: Results from group-based trajectory modeling by Ryota Ishiwata, Abdelrahman AlAshqar, Mariko Miyashita-Ishiwata and Mostafa A Borahay in Women’s Health
Footnotes
Appendix
Covariate, hysterectomy, and medication codes.
| Age |
| Comparison of birth date to hysterectomy procedure |
| Facility type |
| Marketscan field “STDPLAC”: Hospital inpatient, hospital outpatient, or ambulatory surgical center |
| Region |
| Marketscan field “REGION”: Northeast, north central, south, west, and unknown. |
| Insurance |
| Marketscan field “PLANTYP”: Preferred provider organization, comprehensive, exclusive provider organization, health maintenance organization, point of service, point of service with capitation, consumer-driven health plan, high deductible health plan, and missing or unknown |
| Urban or non-urban |
| Marketscan field “MSA”: 0 or 1 |
| Psychiatric diagnosis |
| Mood |
| ICD-9: 29383, 29600, 29601, 29602, 29603, 29604, 29605, 29606, 29610, 29611, 29612, 29613, 29614, 29615, 29616, 29620, 29621, 29622, 29623, 29624, 29625, 29626, 29630, 29631, 29632, 29633, 29634, 29635, 29636, 29650, 29651, 29652, 29653, 29654, 29655, 29656, 29690, 29699, 3004, 311 |
| ICD-10: F0631, F0632, F0633, F0634, F3010, F3011, F3012, F3013, F302, F303, F304, F308, F309, F3130, F3131, F3132, F314, F315, F319, F320, F321, F322, F323, F324, F325, F328, F329, F330, F331, F332, F333, F3340, F3341, F3342, F338, F339, F340, F341 |
| Anxiety |
| ICD-9: 29384, 30000, 30001, 30002, 30009, 30010, 30020, 30021, 30022, 30023, 30029 |
| ICD-10: F064, F4000, F4001, F4002, F4010, F4011, F40210, F40218, F40220, F40228, F40230, F40231, F40232, F40233, F40240, F40241, F40242, F40243, F40248, F40290, F40291, F40298, F408, F409, F410, F411, F413, F418, F419 |
| Insomnia |
| ICD-9: 78050, 78051, 78052 |
| ICD-10: G4700, G4701, G4709 |
| Schizophrenia |
| ICD-9: 29500, 29501, 29502, 29503, 29504, 29505, 29510, 29511, 29512, 29513, 29514, 29515, 29520, 29521, 29522, 29523, 29524, 29525, 29530, 29531, 29532, 29533, 29534, 29535, 29540, 29541, 29542, 29543, 29544, 29545, 29550, 29551, 29552, 29553, 29554, 29555, 29560, 29561, 29562, 29563, 29564, 29565, 29570, 29571, 29572, 29573, 29574, 29575, 29580, 29581, 29582, 29583, 29584, 29585, 29590, 29591, 29592, 29593, 29594, 29595 |
| ICD-10: F200, F201, F202, F203, F205, F2081, F2089, F209, F21, F250, F251, F258, F259 |
| Bipolar disorder |
| ICD-9: 29640, 29641, 29642, 29643, 29644, 29645, 29646, 29660, 29661, 29662, 29663, 29664, 29665, 29666, 2967, 29680, 29681, 29689 |
| ICD-10: F310, F3110, F3111, F3112, F3113, F312, F3160, F3161, F3162, F3163, F3164, F3170, F3171, F3172, F3173, F3174, F3175, F3176, F3177, F3178, F3181, F3189, F319 |
| Psychotic disorders (Paranoid states) |
| ICD-9: 29381,29382 2970, 2971, 2972, 2973, 2978, 2979, 2980, 2981, 2982, 2983, 2984, 2988, 2989 |
| ICD-10: F22, F23, F24, F28, F29, F060, F062 |
| Pain diagnosis |
| Fibromyalgia |
| ICD-9: 7291 |
| ICD-10: M609, M791, M797 |
| Neck pain |
| ICD-9: 7231, 7210X, 7211X, 7220X, 72231, 72.71, 72281, 72291, 723XX, 8390, 8391, 8470 |
| ICD-10: M542, M5481, M4322, M4602, M4642, M4682, M4692, M47022, M4712, M4722, M47812, M47892, M4802, M4812, M4822, M4832, M49832, M5000, M5001, M5002, M50020, M50021, M50022, M50023, M5003, M5010, M5011, M5012, M50120, M50121, M50122, M50123, M5013, M5020, M5021, M5022, M50220, M50221, M50222, M50223, M5023, M5030, M5031, M5032, M50320, M50321, M50322, M50323, M5033, M5090, M5081, M5082, M50820, M50821, M50822, M50823, M5083, M5090, M5091, M5092, M50920, M50921, M50922, M5093, M532X2, M5382, M5412, M542 |
| Back pain |
| ICD-9: 7245, 72230, 72232, 72233, 72270, 72272, 72273, 72280, 72282, 72283, 72290, 72292, 72293, 7384, 8054, 8058, 8392, 83942, 846, 8460, 8471, 8472, 8473, 8479 |
| ICD-10:M545, M546, M5489, M549, M62830 |
| Anemia |
| ICD-9: 2800, 2808, 2809, 2819, 2850, 2851, 28521, 28522, 28529, 2853, 2858, 2859 |
| ICD-10: D508, D509, D539, D649 |
| Menopausal disorders |
| ICD-9:25631 25639 6270 6271 6272 6273 6274 6278 6279 V074 |
| ICD-10: E28310, E28319, E2839, N924, N950, N951, N952, N958, N959 |
| Surgical indication |
| Benign neoplasm of the uterus |
| ICD-9: 2180, 2181, 2182, 2189, 2190, 2191, 2198, 2199 |
| ICD-10: D250, D251, D252, D259, D260, D261, D267, D269 |
| Inflammatory disease of female pelvic organs |
| ICD-9: 6140, 6141, 6142, 6143, 6144, 6145, 6146, 6147, 6148, 6149, 6150, 6151, 6159, 6160, 61610, 61611, 6162, 6163, 6164, 61650, 61651, 6168, 61681, 61689, 6169, 62571 |
| ICD-10: N7001, N7002, N7003, N7011, N7012, N7013, N7091, N7092, N7093, N710, N711, N719, N72, N730, N731, N732, N733, N734, N735, N736, N738, N739, N74, N750, N751, N758, N759, N760, N761, N762, N763, N764, N765, N766, N7681, N7689, N770, N771 |
| Endometriosis |
| ICD-9: 6170, 6171, 6172, 6173, 6174, 6175, 6176, 6178, 6179 |
| ICD-10: N800, N801, N802, N803, N804, N805, N806, N808, N809 |
| Prolapse of female genital organs |
| ICD-9: 6180, 61800, 61801, 61802, 61803, 61804, 61805, 61809, 6181, 6182, 6183, 6184, 6185, 6186, 6187, 6188, 61881, 61882, 61883, 61884, 61889, 6189 |
| ICD-10: N810, N8110, N8111, N8112, N812, N813, N814, N815, N816, N8181, N8182, N8183, N8184, N8185, N8189, N819 |
| Menstrual disorders |
| ICD-9: 6253, 6260, 6261, 6262, 6263, 6264, 6265, 6266, 6268, 6269 |
| ICD-10: N897, N910, N911, N912, N913, N914, N915, N920, N921, N922, N923, N925, N926, N944, N945, N946 |
| Surgical complication |
| Acute renal failure |
| ICD-9: 5845, 5846, 5847, 5848, 5849 |
| ICD-10: N170, N171, N172, N178, N179, N990 |
| Postoperative infection |
| ICD-9: 9985, 99851, 99859 |
| ICD-10: T814XXA, T814XXD, T814XXS |
| Sepsis/shock |
| ICD-9: 31, 362, 380, 3810, 3811, 3812, 3819, 382, 383, 3840, 3841, 3842, 3843, 3844, 3849, 389, 99591, 99592, 63450, 63451, 63452, 63550, 63551, 63552, 63650, 63651, 63652, 63751, 63752, 6385, 6395, 78550, 78551, 78552, 78559, 9980, 99801, 99809 |
| ICD-10: A021, A327, A394, A400, A401, A403, A408, A409, A4101, A4102, A411, A412, A413, A414, A4150, A4151, A4152, A4153, A4159, A4181, A4189, A419, A427, B377, R6520, R6521, R7881, R570, R571, R578, R579, T8110XA, T8110XD, T8110XS, T8111XA, T8111XD, T8111XS, T8112XA, T8112XD, T8112XS, T8119XA, T8119XD, T8119XS, T82211A, T82211D, T82211S, T82212A, T82212D, T82212S, T82213A, T82213D, T82213S, T82218A, T82218D, T82218S, T82221A, T82221D, T82221S, T82222A, T82222D, T82222S, T82223A, T82223D, T82223S, T82228A, T82228D, T82228S, T882XXA, T882XXD, T882XXS |
| Ureteral injury |
| ICD-9: 8672, 8673 |
| ICD-10: S3710XA, S3710XD, S3710XS, S3712XA, S3712XD, S3712XS, S3713XA, S3713XD, S3713XS, S3719XA, S3719XD, S3719XS |
| Urinary tract infection |
| ICD-9: 5990 |
| ICD-10: N390 |
| Wound disruption |
| ICD-9: 9983, 99830, 99831, 99832 |
| ICD-10: T8130XA, T8130XD, T8130XS, T8131XA, T8131XD, T8131XS, T8132XA, T8131XS |
| Hysterectomy |
| Abdominal |
| CPT: 58150, 58152, 58180 |
| Laparoscopic or laparoscopic assisted |
| CPT: 58541, 58542, 58543, 58544, 58550, 58552, 58553, 58554, 58570, 58571, 58572, 58573 |
| Vaginal |
| CPT: 58260, 58262, 58263, 58267, 58270, 58275, 58280, 58290, 58291, 58292, 58293, 58294 |
| Obstetric delivery |
| CPT: 59400, 59409, 59410, 59412, 59510, 59514, 59515, 59610, 59612, 59614, 59618, 59620 ICD 10: O80, O82 |
| List of antidepressants |
| Class |
| Generic names |
| Selective serotonin reuptake inhibitor |
| Citalopram |
| Escitalopram |
| Fluoxetine |
| Fluvoxamine |
| Paroxetine |
| Sertraline |
| Serotonin-norepinephrine Reuptake Inhibitor (SNRI) |
| Desvenlafaxine |
| Duloxetine |
| Venlafaxine |
| Milnacipran |
| Levomilnacipran |
| Benzodiazepines |
| Lorazepam |
| Alprazolam |
| Buspirone |
| Diazepam |
| Chlordiazepoxide |
| Midazolam |
| Antidepressants |
| Bupropion |
| Vilazodone |
| Vortioxetine |
| Anticonvulsant |
| Aripiprazole |
| Asenapine |
| Carbamazepine |
| Lamotrigine |
| Lurasidone |
| Olanzapine |
| Quetiapine Fumarate |
| Ziprasidone |
| TCAs/TeCAs |
| Amitriptyline |
| Amoxapine |
| Clomipramine |
| Desipramine |
| Desonide |
| Doxepin |
| Imipramine |
| Maprotiline |
| Mirtazapine |
| Nortriptyline |
| Protriptyline |
| Trimipramine maleate |
| Other |
| Lithium |
| Lithium |
| Monoamine oxidase inhibitors |
| Selegiline |
| Isocarboxazid |
| Phenelzine |
| Tranylcypromine |
CPT: Current Procedural Terminology; ICD-CM: International Classification of Disease, Clinical Modification.
Acknowledgements
We would like to thank Dr. Keisuke Tada for valuable feedbacks to enhaunce the quality of this research project.
Declarations
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
