Wedging Doors Closed

Q: Is it allowed to use door wedges to keep the exam doors locked? If an intruder enters our hospital, there was some talk about using the door wedges to lock the doors so the intruder cannot open the doors. It was also brought up that if one person was in the room and used a door wedge and had a medical emergency, the door could not be opened to help the person. Any help you could give us would be greatly appreciated?

A: During normal operations, it would not be acceptable to wedge a door closed to an exam room or a patient sleeping room as that would cause an unsafe environment, and would likely be cited under CMS Condition of Participation standard §482.41(a) for an unsafe environment. Your intuition is correct: A wedged door would cause delay in gaining access to a patient in distress.

However, during an emergency, all “bets are off”, meaning you do what you have to do to respond to the course of the emergency. If this means you wedge the door closed to prevent an intruder from entering the room, then that’s what you do. Although you won’t find this written in any code or standard, the concept of emergency response is you do whatever is needed to provide care and safety for your patients. Wedging a door closed to prevent an intruder from entering the room would be an acceptable plan in my book. You just don’t do that during normal operations.

Locks on Operating Room Doors

Q: Our operating rooms were constructed with deadbolt locks on each surgical room. I’m told they did this to keep someone from entering the room during a case. Since this is most likely the path of egress I can’t see this being okay by the LSC. Your thoughts?

A: I would agree with you… But there are quite of few variables here. First of all, is the OR area a suite of rooms? If so, then the doors to the operating rooms do not have to latch, and a thumb-turn handle on the egress side to unlock a dead-bolt would be permitted since there is only one action to operate the door (see of the 2012 Life Safety Code). But, if the operating rooms are not in a suite, then no, you would not be able to have dead-bolt locks because the doors would have to latch (since they open onto a corridor) and that would mean there would be two actions to operate the door: 1) To unlock the dead-bolt, and 2) Unlatch the latchset hardware.

Now section of the 2012 Life Safety Code modified some of that and says existing doors are permitted to have two releasing operations to a room serving not more than 3 occupants, provided it does not require simultaneous operations to unlock the door. But… the typical operating room serves more than 3 occupants, so I don’t see that section working for you.

I would say it boils down to whether or not the OR is a suite of rooms. If yes, then the dead-bolt locks may stay if there is only one action to operate the door. If no, then you have to remove the dead-bolt locks.

Locked Mechanical Room

Q: The computer server is located in the mechanical room. IT wants the room locked for HIPAA compliance. Is it permissible to lock a mechanical room? Doesn’t locking compromise access to electrical panels, fire panel, water shutoff, etc. in an emergency?

A: Yes… you can lock a door into a mechanical room, as far as NFPA is concerned; you just can’t lock the door in the path of egress. Does a locked door to a mechanical room restrict access to electrical distribution panels? Yes it does, but that is a good thing. According to NFPA 99-2012, section access to overcurrent protective devices (i.e. circuit breakers) serving Category 1 or Category 2 rooms must be secured to allow access for authorized individuals only. So, having these distribution panels inside a locked mechanical room meets the requirement of NFPA 99.

It is expected that authorized individuals will have a key or device to access this room in the event of an emergency. Make sure any locks that are installed on the door does not lock the door for those individuals exiting the mechanical room. It is not uncommon for Information Technology to place intermediate distribution frames containing servers in various locations around the facility. The challenge is to meet all of the requirements for limiting access for HIPAA compliance and still allow regular access for other items in the room.

Dead Bolt Locks on Office Doors

Q: Does section of the 2012 edition of the LSC, which prohibits more than one releasing actions to operate the door apply to office doors within a healthcare occupancy? For example, the nurse manager’s office opens to the corridor, and she wants to place a deadbolt lock on the door. Does the addition of a deadbolt lock create a violation? Does egress from a single office require the same “single motion” requirement as the remainder of the path of egress?

A: Yes, it does. As long as the door (no matter where the door is located) is in the path of egress, then it must comply with and be operable with only one releasing motion. A door to an office qualifies as a door in the path of egress, because if you’re inside that office, the door in in your path to the way to get outdoors. Now, it is possible that if there were two entrances (doors) to the same office, you could designate one of the doors as being in the path of egress, and the other door as not being in the path of egress. This way, the door that is not in the path of egress could have a dead-bolt lock that requires more than one releasing motion, but that may not help you with this situation.

Please understand that since this requirement to have no more than one releasing operation to operate the door is found in chapter 7 of the 2012 LSC, it applies to all occupancies, with the exception of residential occupancies as the standard states. This means it applies to your medical office buildings, administrative buildings, clinics, and as well as your hospital. Deadbolt locks that operate separately from the door latch set just are not permitted.

The 2012 LSC section does allow existing conditions where two releasing operations on a door serving an area having an occupant load not exceeding three persons to remain. But that does not allow you to install locks on that door… it is only available for existing conditions.

Rest assured, there are multiple types of locks that are available for doors that can be operated with only one releasing motion. The most common includes a deadbolt that retracts when the door handle is operated, and those are commonly found in hotels. There are other types of locks that are incorporated into the door handle and will unlock the door when the door handle is turned.

Patient Sleeping Room Locks

Q: Are locks permitted on patient room doors? Where can I locate the NFPA requirements for adding new hardware to patient room doors?

A: Section of the 2012 Life Safety Code says locks are not permitted on patient sleeping room doors. Then, an exception to this standard says key-locking devices that restrict access to the room from the corridor and that are operable only by staff from the corridor side shall be permitted. Such devices shall not restrict egress from the room. What this means is you can lock the door to a patient sleeping room as long as the person on the inside of the room can open the door and get out.

However, before you think about adding deadbolt locks to existing doors, section of the same code says you cannot have more than one lock or latch to operate the door. This means a deadbolt lock that is separate from the door latch set is not permitted because it takes two actions to operate the door: 1) Unlock the lock, and; 2) Turn the latch set handle. What you can have is a lock that automatically unlocks the door when the latch set handle is turned. These are also called hotel suite locks, because they are common in hotels. There is a deadbolt that is integrated with the latch set, and a person may unlock the door by simply grasping the latch set handle and turning.

If by chance the door in question is a fire-rated door, according to NFPA 80 you are permitted to make minor changes to the door in order to install new hardware, provided the hardware is listed for use on a fire rated door assembly.

Dead-bolt Locks on Bathroom Doors

Q: I’ve read from your column that dead-bolt locks are not permitted on doors in hospitals, but I’ve seen in other facilities where dead-bolt locks are installed on bathroom doors. Is this permitted?

A: Actually, dead-bolt locks with a thumb-turn on the inside would be permitted on restroom doors, provided the doors to the restrooms are not positive latching. Restroom doors are not required to latch, therefore the larger restrooms rarely have positive latching hardware. Section of the 2000 Life Safety Code© says doors in the means of egress are not permitted to have more than one action to operate the door. If the door had a latch-set and a dead-bolt lock then that would not be permitted since it takes two actions to operate the door (unlock the dead-bolt and turn the latch-set). But since the bathroom door does not require a latch-set, then a dead-bolt lock that can be unlocked from the egress side would be permitted.

Strange Observations – Part 6

Door with Conflicting SIgnage Web 2Continuing in a series of strange things that I have seen while consulting at hospitals….

This series of pictures is just plain sad… and wrong. The top picture shows a door in the path of egress with an ‘Exit’ sign over it. If you look closely, there is a magnetic lock on this door and there is a sign posted on the door as well. This door is in an exit access corridor and lead from one physician’s office area to another physician’s office area and for some reason was equipped with a magnetic lock. But the magnetic lock did not qualify for the exceptions to of the 2000 LSC for delayed egress locks or access-control locks.

Take a look at the bottom picture… The picture is a close-up of the sign on the door. The sign says “This is not an exit…” but the ‘Exit’ sign over the door says “This is an exit”. That’s a conflict. To be sure, the door is an exit as this door was located in an exit access corridor and without the designated exit it would have been over a 100 foot dead-end corridor (which is not permitted). So, the exit is necessary.

But the conflicting signage is just wrong. Just because one physician’s office did not want people exiting into his area, you cannot lock the door and say the door is not an exit.

Close up of Door With Conflicting Signage Web 2I remember we tried to find out what actually releases the magnetic lock and the staff that I was with could not say. They said they were not sure if it released on a fire alarm and they checked with security and it did not release on any signal that security could transmit.

This is the type of deficiency that will lead to an Immediate Jeopardy decision and cause all sorts of headaches for a hospital.

Strange Observations – Part 4

Deadbolt on Exit Door Web 2Continuing in a series of strange things that I have seen while consulting at hospitals….

This is a picture of a door to an exit stairwell. Note that there is a deadbolt lock on the door above the fire exit hardware. That is so not allowed, according to of the 2000 LSC. You cannot lock a door in the path of egress that requires the use of a tool or a key. (There are some exceptions but this door does not qualify).

Also, the sign on the door indicates it is a delayed egress lock but it says the door can be opened in 30 seconds. Section allows for a 30 second delay but only when it is approved by the authority having jurisdiction (AHJ).  Most national AHJs for hospitals do not approve of the 30 second delay so this is not permitted. (Some hospitals mistakenly believe that the local or state fire marshal is the only AHJ they need to ask permission, but they don’t realize that the accreditation organizations and CMS are AHJs as well.) If you want to utilize an exception in the LSC whereby it is permitted with the AHJ’s permission, then you need to have permission from all the AHJs who have authority over hospitals:


  • CMS (Federal agency who controls Medicare/Medicaid reimbursements)
  • Accreditation organization
  • State agency who license hospitals
  • State fire marshal
  • Local fire inspector
  • Insurance company

The likelihood of getting all of them to agree on an exception to the LSC is slim, if not impossible.

Controlled Access Locks

Q: While conducting fire drills in the hospital, one of the questions on our drill evaluation sheet is, “Did the security doors in the fire zone release properly?” We have controlled areas where the doors are locked to control access into the unit. To exit the unit only requires the push of a button and the doors release. So are we in compliant with this controlled access not releasing during the fire alarm activation since the exit is not controlled? Or should the doors release to allow free entry and exits?

A: Doors in the path of egress in a healthcare occupancy are not permitted to be locked. However, there are three (3) exceptions to this requirement:

  • Delayed egress locks complying with section, 2000 LSC
  • Access-control locks complying with section
  • Clinical needs locks complying with section

By the sound of your situation, it appears to me that you do not have delayed egress locks and you do not have clinical needs locks, which leaves access-control locks. However, it also appears that your description of the security door locks may not be in compliance with section Here is a summary of the requirements for access-control locks:

  1. A motion sensor must be mounted on the egress side to detect occupants approaching the door, and automatically unlock the door in the direction of egress
  2. A loss of power to the control system automatically unlocks the door in the direction of egress
  3. A manual release button must be mounted 40 to 48 inches above the floor, and within 5 feet of the door, that when operated will directly interrupt the power to the lock, independent of the control system, for a minimum of 30 seconds. The button must be labeled with the words “PUSH TO EXIT”.
  4. The door must unlock in the direction of egress upon activation of the building fire alarm system or the building sprinkler system.

So, it appears to me that you are missing the motion sensor on the egress side of the door that would automatically unlock the door when someone approaches. Also, it sounds like your locks are not interconnected to the building fire alarm system to automatically unlock on an alarm. According to section, these are required. Also, check the ‘PUSH TO EXIT” button to make sure it interrupts power to the locks for a minimum of 30 seconds, when depressed.

Expert Witness

imagesXLH1EQVPI received a telephone call recently from a lawyer that asked me if I would be willing to serve as an expert witness in a pending litigation suit. The situation of the lawsuit is a patient who was admitted to the inpatient psychiatric unit of a hospital busted the protective cover over the fire alarm manual pull station, and activated the fire alarm system. The entrance doors to the psychiatric unit unlocked on the fire alarm, and the patient ran out, climbed to the roof of the facility and jumped off.

Now the family of the patient is suing the hospital and the fire alarm company who installed the interface between the door locks and the fire alarm system. The plaintiff’s lawyer is arguing that the interface should never have been installed and the doors should not have unlocked upon activation of the fire alarm system. I agree with the plaintiff’s lawyer, but the problem is, it was the defense’s lawyer who contacted me.

The two major codes and standards that apply in this situation are the NFPA 72 National Fire Alarm Code (1999 edition), and the NFPA 101 Life Safety Code (2000 edition). The Life Safety Code (LSC) allows clinical needs locks on doors in the path of egress, but only in healthcare occupancies (hospitals). They are not permitted in any other occupancy. These types of locks are permitted where the “clinical needs of the patients require specialized security measures for their safety, provided that staff can readily unlock such doors at all times”. Clinical needs locks are permitted for Behavioral Health units, such as psychiatric and Alzheimer units, to prevent patients from leaving the unit unauthorized, and potentially harming themselves or others.  Clinical needs locks requires all staff who work on that unit (which includes physicians, nurses, aids, clerical, maintenance, foodservice, housekeeping, etc.) to have a key or device to unlock the door on their person at all times. [Code reference:, Exception No. 1, 2000 edition of the LSC].

But when it comes to connecting those clinical needs locks to the fire alarm system so they automatically unlock on a fire alarm activation, the codes are not so clear. Other permissible locks, such as delayed egress locks and access control locks, are required by code to be connected to the fire alarm system and automatically unlock the door whenever there is an alarm. However, there is no such similar language for clinical needs locks to unlock on an alarm. In addition, NFPA 72 section 3-9.7.1 and 3-9.7.2 says if you have a lock on an exit door it must be connected to the fire alarm system, and it must unlock the door on a fire alarm. However, the exception to 3-9.7.2 essentially says this is not required if an AHJ says so, or if another code says so. So, we go back to the Life Safety Code, and we see that delayed egress locks and access control locks are required to unlock on an alarm, but the section on clinical needs locks is very silent on the subject. When a code is silent on a subject that means it is open for interpretation. Therefore, section 3-9.7 of NFPA 72 does not apply to clinical needs locks in a hospital, because the interpretation Life Safety Code permits it. It is documented that psychiatric patients are astute enough to actuate a fire alarm to unlock the exit doors. Therefore, the code is available to be interpreted in such a way to not require clinical needs locks to unlock the doors in the path of egress on activation of the fire alarm system provided all staff carry a key to unlock the door in case of an emergency.

But who makes that interpretation? Usually the authorities having jurisdiction (AHJ) makes that interpretation. Section of the 2000 LSC says the AHJ shall determine whether the provisions of the LSC are met. But when the AHJ does not issue an interpretation, then it is up to the individual organization’s to decide. As far as I know, there is no national AHJ for healthcare that has it written in their standards that clinical needs locks should not be connected to the fire alarm system. Therefore, the hospital needs to ask the question: Based on the potential risk to patient safety, is it best to connect the door locks on the psychiatric unit to the fire alarm system so they unlock upon activation of the fire alarm system?

I declined the offer to be an expert witness for the defense, and suggested they settle the lawsuit out of court.