An Electronic Medical Record–Based Decision Support Tool for the Diagnosis of Primary Hyperparathyroidism

Primary hyperparathyroidism (PHPT), the predominant cause of chronic hypercalcemia, is diagnosed by elevated serum calcium levels in the presence of elevated or inappropriately normal parathyroid hormone (PTH) levels.¹ Complications of untreated PHPT include osteoporosis, fragility fractures, and nephrolithiasis.² Definitive treatment is achieved with surgical parathyroidectomy.³ Studies have reported that PHPT is underdiagnosed and undertreated and have recommended development of clinical decision support tools to aid in PHPT management.⁴ This cohort study evaluated the association of an electronic medical record (EMR)–based tool with frequency of serum PTH level measurement among patients with chronic hypercalcemia and described their subsequent treatment in a large academic health system.

The EMR-based tool (eFigure in the Supplement) identified outpatients with chronic hypercalcemia from the EMR of UCLA Health and triggered an alert prompting physicians to order a PTH level measurement. This study followed the STROBE reporting guideline and was approved by the UCLA institutional review board, which waived informed consent because data were deidentified.

Patients were identified from May to November 2020 and were followed up for 16 months. The alert algorithm was triggered in adults (aged ≥18 years) with 2 instances of hypercalcemia (total calcium level >10.4 mg/dL [to convert to mmol/L, multiply by 0.25]) in a 6-month interval and absent PTH level measurement in the following year. The tool encoded logic to exclude patients with secondary hyperparathyroidism, tertiary hyperparathyroidism, and nonparathyroid causes of hypercalcemia (eTables 1 and 2 in the Supplement). Baseline sequelae of PHPT were evaluated using ICD-9 and ICD-10 codes (eTable 3 in the Supplement). Patients with confirmed hypercalcemia and an elevated or inappropriately normal serum PTH level (>11 pg/mL [to convert to ng/L, multiply by 1]) were diagnosed with PHPT based on biochemical findings. During follow-up, we examined the frequencies of bone mineral density assessment by dual energy x-ray absorptiometry, kidney ultrasonography, parathyroidectomy, and initiation of pharmacologic therapy for osteopenia or osteoporosis (eMethods in the Supplement). Two-sided P < .05 was considered significant. We analyzed data using R (version 3.6.2).

The population included 395 patients with chronic hypercalcemia (Table). During the baseline period, 9 patients (7%) who would have met alert criteria subsequently had their serum PTH measured. This rate increased to 45% (n = 31) during the patient identification period (Figure) and 54% (n = 215) during 6-month follow-up; 213 (99%) were diagnosed with PHPT, but 13 (6%) had low to normal PTH levels (11-20 ng/L). Physicians’ responses to the tool were as follows: 182 patients (46%) had PTH ordered, 87 (22%) alerts were dismissed, 59 (15%) were deferred to another practitioner, and 67 (17%) were deemed inappropriate for the encounter.

At 16 months’ follow-up, 79 patients (37%) with PTH assessment had a bone mineral density scan ordered vs 22 of 180 (12%) without PTH assessment (P < .001). Thirty-six patients (17%) with PTH assessment were given new diagnoses of osteopenia or osteoporosis vs 7 (4%) without PTH assessment (P < .001). Parathyroidectomy was performed in 21 patients (10%) with PTH assessment vs 0 without PTH assessment (P < .001). Kidney ultrasonography was ordered for 34 patients (16%) with PTH assessment vs 18 (10%) without PTH assessment (P = .10).

The EMR-based intervention was associated with a 6-fold increase in serum PTH assessment among outpatients with chronic hypercalcemia, nearly all of whom had PHPT. After PTH assessment, the frequency of appropriate downstream diagnostic and therapeutic interventions increased in the absence of further prompting. The EMR-based tool extracted abnormal laboratory results from a patient’s EMR and nudged physicians toward appropriate diagnostic action, while still preserving their decision-making autonomy.⁵ The tool underwent prelaunch refinements to minimize disruptions to physician workflow and reduce cognitive strain.

Limitations include reliance on diagnosis and procedure codes for data extraction, examination of total serum rather than albumin-adjusted or ionized calcium levels, and inclusion of patients with low to normal PTH levels. This study suggests a role for implementation science methods as a solution to PHPT underdiagnosis and undertreatment.

Accepted for Publication: May 18, 2022.

Published Online: July 25, 2022. doi:10.1001/jamainternmed.2022.2684

Corresponding Author: Michael W. Yeh, MD, Section of Endocrine Surgery, Department of Surgery, UCLA David Geffen School of Medicine, 100 UCLA Medical Plaza, Ste 310, Los Angeles, CA 90095 (MYeh@mednet.ucla.edu).

Author Contributions: Mss Dawood and Shu contributed equally to the study. Mss Dawood and Shu had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Dawood, Shu, Leung, Yeh.

Acquisition, analysis, or interpretation of data: Dawood, Shu, Tseng, Kim, Nguyen, Yeh.

Drafting of the manuscript: Dawood, Shu, Kim, Nguyen, Yeh.

Critical revision of the manuscript for important intellectual content: Dawood, Shu, Tseng, Leung, Yeh.

Statistical analysis: Dawood, Shu, Tseng, Kim, Nguyen.

Obtained funding: Yeh.

Administrative, technical, or material support: Dawood, Nguyen, Yeh.

Supervision: Leung, Yeh.

Conflict of Interest Disclosures: None reported.