ISMP Survey Shows Provider Text Messaging Often Runs Afoul of Patient Safety
In June through August 2017, we invited readers of our acute care, community/ambulatory care, long-term care, and nursing newsletters to complete an online survey about the texting of medical orders in healthcare. We sincerely thank the 778 respondents who completed the survey, which included nurses (40%, n=312), pharmacists (38%, n=299), physicians and other prescribers (7%, n=54), medication/patient safety officers and quality/risk managers (7%, n=53), and others (8%, n=60, educators, pharmacy technicians, etc.). Almost all respondents were from the US (95%) and practiced in a hospital setting (86%). About two-thirds (63%) of the respondents were staff-level practitioners, and the remaining were managers (21%), directors (10%), or administrators (6%).
ISMP conducted this survey to gain insight into the practice of texting medical orders given the ongoing debate regarding its use.1 Technology-savvy healthcare professionals have embraced the convenience of this 21st century form of communication, while opponents feel it is too informal to properly document patient care. They also have concerns about data security and the potential impact on patient safety with texting medical orders. Both sides of the debate offer compelling viewpoints, which were both clearly evident in the survey results of US respondents (n=742) that follow.
Survey Results: Scope of Texting in Healthcare
Texting opinions. Thirty-three percent of all respondents, and more than half (55%) of all medication/patient safety officers and risk/quality managers, do not believe medical orders should be texted under any circumstances in healthcare. Another 40% of all respondents thought the practice was acceptable only when using an encrypted device application (e.g., TigerText, Doc Halo). While 26% of physicians reported that texting should be allowed in any circumstance, only 15% of nurses and pharmacists, and only 4% of medication/patient safety officers and risk/quality managers felt the same way.
Texting policies and practices. More than half (53%) of all respondents told us that texted medical orders are not allowed per policy in their facility. Most of the other respondents said they had no policy on the topic (19%) or were uncertain whether a policy existed (16%). Only 12% of all respondents reported that texting was allowed in their facility per policy—8% for any orders if using an encrypted device, 3% under any circumstances, and 1% under certain circumstances (e.g., to communicate information other than medication orders, for clarifications only, to alert prescribers to call).
Still, 45% of pharmacists and 35% of nurses reported that medical orders are regularly being texted irrespective of a policy. Another 36% are uncertain whether the practice occurs. No physicians reported awareness of a policy that allowed the practice, yet 38% reported that medical orders are being texted. Less than a quarter (22%) of respondents were certain that medical orders are not being texted in their facility.
Frequency of texting. Among respondents who reported receiving texted orders during the past year, more than half receive them every day (20%) or at least every week (35%). Another 17% receive texted orders once or twice a month, and the remainder (28%) receive them less frequently. Pharmacists reported receiving texted orders more frequently than nurses.
Types of texted orders. For those who thought texted orders should be allowed in healthcare (67%), half thought orders for chemotherapy (50%) and complex order sets (54%) such as those for parenteral nutrition should be prohibited. More than a quarter thought orders for high-alert medications (30%) and controlled substances (29%) should not be allowed. However, in practice, very few types of texted orders were prohibited in respondents’ facilities. Only 9% said that texted chemotherapy orders were not allowed, and only 3% reported that complex order sets could not be texted. Less than 2% reported any further restrictions, although more than half (56%) of the respondents said they were uncertain about restrictions or were unaware of any policy on the topic.
Devices for texting. Among facilities where texted orders have been received during the past year, more than two-thirds (69%) reported that the facility allows the use of standard cell phones. In fact, about 42% indicated that standard cell phones are the only device from which texted orders have been received in the past year. Approximately half (48%) of all respondents told us that they have received texted orders via an encrypted device during the past year, but only 24% of all respondents reported that this is the only way they have received texted orders. About a quarter (25%) of all respondents said they have received texted orders from both a standard cell phone and an encrypted device during the past year. Thus, even when encrypted devices are available, standard cell phones are still being used to send and receive texted orders.
Texting to clarify orders. More than half of all respondents (55%) have sent text messages to prescribers to ask questions or to clarify orders that may be unclear, incorrect, or inappropriate. This practice was more frequently reported by pharmacists (65%) and physicians (62%) than by nurses (47%). Almost all respondents (86%) who sent text messages to prescribers to clarify orders reported that prescribers responded or replied to these messages via texting.
Documenting texted orders. According to nurses and pharmacists, texted orders are almost always (98%) entered into the patient’s medical record by the person receiving it, similar to a verbal order. However, numerous respondents commented that the order may not be specifically identified as a texted order. Less than 2% of respondents reported that the texted order is automatically entered into the health record by the technology being used.
Survey Results: Risks with Texting Orders
Most respondents reported a high level of concern regarding potential risks associated with the texting of medical orders (Table 1). Overall, medication/patient safety officers and risk/quality managers and nurses were more concerned about the potential risks associated with texted orders, and physicians were least concerned about the risks. Respondents reported that the five most concerning risks associated with texted orders were associated with safety issues impacting order clarity, completeness, and correctness, rather than information security, authentication, or documentation issues:
- Unintended phone/device autocorrection. The majority of all respondents were concerned (15%) or highly concerned (55%) about the risk of unintended autocorrection of medical terms, abbreviations, drug names, or patient names since they are unlikely to be in the device’s dictionary. This could lead to incorrect entries which, if unnoticed by the prescriber or other practitioners, could lead to a delay in care if the order must be clarified, or to a clinically significant error. Most comments from respondents indicated that any autocorrection feature should be disabled on devices used for texting orders to prevent inaccurate corrections since most people fail to reread messages before sending.
- Use of potentially confusing abbreviated text terminology (e.g., 2day). Nearly two-thirds of all respondents were concerned (16%) or highly concerned (50%) about the risk of using abbreviated text terminology (e.g., 2day for today, 2 for to, b/4 for before, 3D for 3 times daily, MT for empty). Nurses and pharmacists reported that about 19% of the texted orders they had received in the past year contained abbreviated text terminology, but such occurrences were infrequent. However, almost half (46%) of all physician respondents reported that texted orders contained these potentially confusing abbreviations, and 30% said it happened frequently in more than a quarter of all texted orders.
- Potential for patient misidentification. A majority of all respondents were concerned (19%) or highly concerned (41%) about the risk of misidentifying the patient with a texted order since most transmission devices and phones may not facilitate the communication of two unique patient identifiers. Medications could be dispensed and administered to the wrong patient if a spelling error occurs, or autocorrection changes the intended patient’s name. In addition, there were repeated comments from respondents who alarmingly said they only include the patient’s initials, unit, room number, or another abbreviated patent identifier with their text messages and orders to offset the risks associated with the security of protected patient information.
- Misspellings. Well more than half of all respondents were concerned (17%) or highly concerned (41%) about the risk of spelling errors with patient names or drugs (particularly similar drug names) and doses. Most texted orders must be entered as free-text rather than selecting drugs and doses from a drop-down menu, or via a voice-recognition feature that may mishear and, thus, misspell words, including drug names.
- Incomplete orders. More than half of all respondents were concerned (20%) or highly concerned (36%) about the risk of communicating incomplete orders when texting. Free-text orders that lack the prompts often found in electronic prescribing systems may be missing critical components, such as the route of administration or, for pediatric weight-based medications, the mg/kg dose. Having no way to prevent an order from being sent via text without all the required components, the need for more clarifications than with electronic prescribing systems, and issues with punctuation and hitting “send” before the order has been completed, were also frequent concerns listed by respondents.
Half or more of all respondents felt that these risks, along with the failure to retain or document the text message, and the inability to authenticate the sender/receiver, were of the highest concern. The least concerning risks, which still clearly worried most respondents, included a delay in receipt or transcription of texted orders, and the lack of prescriber clinical decision support while texting. For the latter concern, many respondents commented that nurses who enter the texted order into the prescribing system, and pharmacists who verify the order, should receive any alerts and clarify the orders if concerns arise. However, other respondents commented that the lack of decision support when prescribing could lead to unnecessary variation in practices and transfer responsibility for the correctness of the order from the prescriber to the nurse (or pharmacist).
Survey Results: Errors with Texted Orders
Seven percent of all respondents were aware of errors or close calls that have occurred involving a texted order. While this does not seem to be an excessive amount of errors, those described by respondents were primarily associated with the set of risks that were described above, some of which are unique to texted orders (Table 2).
Given that texting is just too convenient, many in healthcare feel that the text messaging of orders is unlikely to go away, despite policies prohibiting their use or the known safety concerns. Our survey results tend to support this conclusion, although more scientifically rigorous research should be conducted to further confirm the scope of current use. The benefits of texting orders are primarily related to its popularity and convenience, workflow synergy and speed, and perception of similar risks when compared to other forms of communicating orders. In fact, numerous respondents to our survey noted that, while texting of medical orders is clearly not as safe as electronic prescribing, it may be safer and more timely than verbal or telephone orders.
We disagree, particularly given that verbal or telephone orders can be read back to ensure accuracy and understanding, and because most practitioners who responded to our survey are texting orders via standard cell phones or devices without encryption or critically important safety features. While other forms of communicating medical orders carry some of the same risks as texting orders, the informal nature of texting orders, often without a known policy or procedure associated with the process, has resulted in uniquely alarming risks, including abbreviated language, improper autocorrection, and texting orders without full patient names and a second unique identifier to offset some data security concerns, to name a few.
The texting of medication-specific orders should not be allowed until the safety issues have been identified and resolved through advanced technology along with the development of vetted, industry-wide clinical guidelines that can be employed in organizations to ensure standardized, safe, and secure texting processes. Leadership must establish and communicate policies on the texting of orders and take a strong stance on avoiding texted medication-specific orders at this time until they can be safely introduced into healthcare through careful pilot testing and implementation plans.
- ISMP. The texting debate: beneficial means of communication or safety and security risk? ISMP Medication Safety Alert! 2017;22:1-4