Decreased Utilization of Low-Value Health Care Services during the COVID-19 Pandemic
Tracy Xuan Han, MD, MBA1, David-Dan Nguyen, MPH1, Peter Herzog, MPH1, Stuart Lipsitz, ScD1, Alexander P. Cole, MD1, Tony K. Choueiri, MD2, Quoc-Dien Trinh, MD, MBA1.
1Brigham and Women's Hospital, Boston, MA, USA, 2Dana–Farber Cancer Institute, Boston, MA, USA.
BACKGROUNDLow-value health care is defined as care in which the potential to cause harm is greater than the potential benefits. We hypothesize that rationing of health care during the pandemic decreased the delivery of low-value services.
METHODSData was retrieved from the Research Patient Data Registry and approved by Mass General Brigham Institutional Review Board.
High-value care services were defined according to USPTF guidelines, while low-value care services were selected based on previous literature. The periods considered were of 3 months each and consisted of the study period during the pandemic year (Q4: March 2, 2020 to June 1, 2020) and periods preceding the onset of the pandemic (Q1: December 1, 2018 to March 1, 2019; Q2: March 2, 2019 to June 1, 2019; and Q3: December 1, 2019 to March 1, 2020). The numbers of services during each period (NQ) were evaluated.
The 2020 ratio (Y2020 = NQ4/NQ3) represents the change due to the pandemic. To account for seasonality, the 2019 ratio in number of services (Y2019 = NQ2/NQ1) was used to illustrate relative service counts during a typical year. Y2020-Y2019 less than zero reflects a reduction in the service during the pandemic year accounting for seasonality. The calculation was made for each service as well as by value of care and by type of care (cancer vs. non-cancer). The difference between YLow-value and YHigh-value is the difference-in-differences (DID) of ratios and illustrates the differential decline in services by value of care during the pandemic period. YHigh-value- YLow-value greater than zero suggests that low-value care declined to a greater degree than high-value care.
RESULTS
Demographic characteristics of 3,271,957 patients during the four time periods are obtained. Eighteen of the 21 identified services had a reduction in volume during the pandemic period. Reductions of both low-value and high-value care during the pandemic were evident overall. The DID in ratios of high-value care and low-value care was 0.08 (p<0.01), suggesting a modest greater decline in low-value care as compared to high-value care during the pandemic. The reduction in low-value care relative to the decline in high-value care is most pronounced for cancer care with a DID in ratios of 3.39 (p<0.01). All DID of ratios are presented in Table 1.
CONCLUSION There was a reduction in both low-value and high-value care with a greater reduction in low-value services, especially for cancer care. The greater degree of reduction in low-value care and within the subdivision of low-value cancer care promisingly suggests that health care providers have appropriately emphasized higher-value care during the pandemic.
Table 1 - Difference-in-differences estimates of ratios of services delivered between the pre-pandemic period and pandemic period.
| PROCEDURE | Ratio2019 (Q2/Q1) | Ratio2020 (Q4/Q3) | Difference in ratios | P-valuea |
| HIGH-VALUE SERVICES | ||||
| Flu vaccine | 0.343 | 0.194 | -0.149 | <0.01 |
| Tobacco cessation | 2.444 | 1.231 | -1.214 | 0.257 |
| Cervical cancer screen | 1.159 | 0.239 | -0.920 | <0.01 |
| Colorectal cancer screen | 1.181 | 0.247 | -0.935 | <0.01 |
| LOW-VALUE SERVICES | ||||
| PSA testing for men over >75 | 1.101 | 0.293 | -0.808 | <0.01 |
| Bone mineral density testing at frequent intervals | 1.111 | 0.221 | -0.889 | <0.01 |
| Homocysteine testing for CV disease | 1.142 | 0.427 | -0.716 | <0.01 |
| Hypercoagulability testing for patients with DVT | 1.170 | 0.627 | -0.543 | <0.01 |
| CT of sinuses for uncomplicated acute rhinosinusitis | 1.014 | 0.336 | -0.678 | <0.01 |
| Head imaging in evaluation of syncope | 0.826 | 0.465 | -0.362 | 0.002 |
| Head imaging for uncomplicated headache | 1.525 | 0.527 | -0.998 | <0.01 |
| EEG for headaches | 0.727 | 1.000 | 0.273 | 0.704 |
| Carotid artery disease screening in asymptomatic adults | 1.175 | 0.494 | -0.681 | <0.01 |
| Carotid artery disease screening for syncope | 1.078 | 0.330 | -0.748 | <0.01 |
| Stress testing for stable coronary disease | 1.147 | 0.378 | -0.769 | <0.01 |
| Carotid endarterectomy in asymptomatic patients | 1.308 | 0.162 | -1.146 | <0.01 |
| Arthroscopic surgery for knee osteoarthritis | 0.714 | 0.333 | -0.381 | 0.303 |
| Cervical cancer screen (by age) | 10.027 | 0.216 | -9.811 | <0.01 |
| Cervical cancer screen with ESRD | 1.529 | 0.190 | -1.339 | <0.01 |
| Colorectal cancer screen in elderly patients | 4.254 | 0.400 | -3.854 | <0.01 |
| Colorectal cancer screen in ESRD | 2.067 | 0.583 | -1.483 | 0.036 |
| VALUE OF ALL SERVICES | ||||
| Difference in ratios of low-value care | 1.177 | 0.353 | -0.824 | <0.01 |
| Difference in ratios of high-value care | 0.974 | 0.231 | -0.744 | <0.01 |
| Difference-in-differences of ratios | - | - | 0.080 | <0.01 |
| VALUE OF CANCER SERVICES | ||||
| Difference in ratios of low-value cancer care | 4.6117 | 0.2977 | -4.314 | <0.01 |
| Difference in ratios of high-value cancer care | 1.1646 | 0.2406 | -0.924 | <0.01 |
| Difference-in-differences of ratios | - | - | 3.390 | <0.01 |
a Wald statistic p-value for testing that Difference in ratios = 0.
Back to 2022 Abstracts